WARNING:__main__:
*****************************************
Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. 
*****************************************
| distributed init (rank 7): env://, gpu 7
| distributed init (rank 0): env://, gpu 0
| distributed init (rank 5): env://, gpu 5
| distributed init (rank 1): env://, gpu 1
| distributed init (rank 3): env://, gpu 3
| distributed init (rank 4): env://, gpu 4
| distributed init (rank 2): env://, gpu 2
| distributed init (rank 6): env://, gpu 6
Namespace(aa='rand-m9-mstd0.5-inc1', auto_resume=True, batch_size=256, clip_grad=None, color_jitter=0.4, crop_pct=None, cutmix=0.0, cutmix_minmax=None, data_path='/data/benchmarks/ILSVRC2012_LMDB', data_set='IMNET_LMDB', device='cuda', disable_eval=False, dist_backend='nccl', dist_eval=True, dist_on_itp=False, dist_url='env://', distributed=True, drop_path=0.2, enable_wandb=False, epochs=300, eval=False, eval_data_path=None, finetune='', gpu=0, head_init_scale=1.0, imagenet_default_mean_and_std=True, input_size=224, layer_decay=1.0, layer_scale_init_value=1e-06, local_rank=-1, log_dir=None, lr=0.004, min_lr=1e-06, mixup=0.0, mixup_mode='batch', mixup_prob=1.0, mixup_switch_prob=0.5, model='convnext_tiny', model_ema=False, model_ema_decay=0.9999, model_ema_eval=False, model_ema_force_cpu=False, model_key='model|module', model_prefix='', momentum=0.9, nb_classes=1000, num_workers=10, opt='adamw', opt_betas=None, opt_eps=1e-08, output_dir='./checkpoint', pin_mem=True, project='convnext', rank=0, recount=1, remode='pixel', reprob=0.25, resplit=False, resume='', save_ckpt=True, save_ckpt_freq=1, save_ckpt_num=3, seed=0, smoothing=0.1, start_epoch=0, train_interpolation='bicubic', update_freq=2, use_amp=True, wandb_ckpt=False, warmup_epochs=20, warmup_steps=-1, weight_decay=0.05, weight_decay_end=None, world_size=8)
Transform = 
RandomResizedCropAndInterpolation(size=(224, 224), scale=(0.08, 1.0), ratio=(0.75, 1.3333), interpolation=bicubic)
RandomHorizontalFlip(p=0.5)
RandAugment(n=2, ops=
	AugmentOp(name=AutoContrast, p=0.5, m=9, mstd=0.5)
	AugmentOp(name=Equalize, p=0.5, m=9, mstd=0.5)
	AugmentOp(name=Invert, p=0.5, m=9, mstd=0.5)
	AugmentOp(name=Rotate, p=0.5, m=9, mstd=0.5)
	AugmentOp(name=PosterizeIncreasing, p=0.5, m=9, mstd=0.5)
	AugmentOp(name=SolarizeIncreasing, p=0.5, m=9, mstd=0.5)
	AugmentOp(name=SolarizeAdd, p=0.5, m=9, mstd=0.5)
	AugmentOp(name=ColorIncreasing, p=0.5, m=9, mstd=0.5)
	AugmentOp(name=ContrastIncreasing, p=0.5, m=9, mstd=0.5)
	AugmentOp(name=BrightnessIncreasing, p=0.5, m=9, mstd=0.5)
	AugmentOp(name=SharpnessIncreasing, p=0.5, m=9, mstd=0.5)
	AugmentOp(name=ShearX, p=0.5, m=9, mstd=0.5)
	AugmentOp(name=ShearY, p=0.5, m=9, mstd=0.5)
	AugmentOp(name=TranslateXRel, p=0.5, m=9, mstd=0.5)
	AugmentOp(name=TranslateYRel, p=0.5, m=9, mstd=0.5))
ToTensor()
Normalize(mean=tensor([0.4850, 0.4560, 0.4060]), std=tensor([0.2290, 0.2240, 0.2250]))
RandomErasing(p=0.25, mode=pixel, count=(1, 1))
---------------------------
reading from datapath /data/benchmarks/ILSVRC2012_LMDB
Number of the class = 1000
Transform = 
Resize(size=256, interpolation=bicubic, max_size=None, antialias=None)
CenterCrop(size=(224, 224))
ToTensor()
Normalize(mean=(0.485, 0.456, 0.406), std=(0.229, 0.224, 0.225))
---------------------------
reading from datapath /data/benchmarks/ILSVRC2012_LMDB
Number of the class = 1000
Sampler_train = <torch.utils.data.distributed.DistributedSampler object at 0x7f6084969c40>
Model = MobileNetV3_Small(
  (conv1): Conv2d(3, 16, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
  (bn1): BatchNorm2d(16, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
  (hs1): Hardswish()
  (bneck): Sequential(
    (0): Block(
      (conv1): Conv2d(16, 16, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn1): BatchNorm2d(16, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (act1): ReLU(inplace=True)
      (conv2): Conv2d(16, 16, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), groups=16, bias=False)
      (bn2): BatchNorm2d(16, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (act2): ReLU(inplace=True)
      (se): SeModule(
        (se): Sequential(
          (0): AdaptiveAvgPool2d(output_size=1)
          (1): Conv2d(16, 8, kernel_size=(1, 1), stride=(1, 1), bias=False)
          (2): BatchNorm2d(8, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
          (3): ReLU(inplace=True)
          (4): Conv2d(8, 16, kernel_size=(1, 1), stride=(1, 1), bias=False)
          (5): Hardsigmoid()
        )
      )
      (conv3): Conv2d(16, 16, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn3): BatchNorm2d(16, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (act3): ReLU(inplace=True)
      (skip): Sequential(
        (0): Conv2d(16, 16, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), groups=16, bias=False)
        (1): BatchNorm2d(16, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      )
    )
    (1): Block(
      (conv1): Conv2d(16, 72, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn1): BatchNorm2d(72, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (act1): ReLU(inplace=True)
      (conv2): Conv2d(72, 72, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), groups=72, bias=False)
      (bn2): BatchNorm2d(72, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (act2): ReLU(inplace=True)
      (se): Identity()
      (conv3): Conv2d(72, 24, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn3): BatchNorm2d(24, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (act3): ReLU(inplace=True)
      (skip): Sequential(
        (0): Conv2d(16, 16, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), groups=16, bias=False)
        (1): BatchNorm2d(16, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
        (2): Conv2d(16, 24, kernel_size=(1, 1), stride=(1, 1))
        (3): BatchNorm2d(24, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      )
    )
    (2): Block(
      (conv1): Conv2d(24, 88, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn1): BatchNorm2d(88, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (act1): ReLU(inplace=True)
      (conv2): Conv2d(88, 88, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=88, bias=False)
      (bn2): BatchNorm2d(88, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (act2): ReLU(inplace=True)
      (se): Identity()
      (conv3): Conv2d(88, 24, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn3): BatchNorm2d(24, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (act3): ReLU(inplace=True)
    )
    (3): Block(
      (conv1): Conv2d(24, 96, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn1): BatchNorm2d(96, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (act1): Hardswish()
      (conv2): Conv2d(96, 96, kernel_size=(5, 5), stride=(2, 2), padding=(2, 2), groups=96, bias=False)
      (bn2): BatchNorm2d(96, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (act2): Hardswish()
      (se): SeModule(
        (se): Sequential(
          (0): AdaptiveAvgPool2d(output_size=1)
          (1): Conv2d(96, 24, kernel_size=(1, 1), stride=(1, 1), bias=False)
          (2): BatchNorm2d(24, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
          (3): ReLU(inplace=True)
          (4): Conv2d(24, 96, kernel_size=(1, 1), stride=(1, 1), bias=False)
          (5): Hardsigmoid()
        )
      )
      (conv3): Conv2d(96, 40, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn3): BatchNorm2d(40, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (act3): Hardswish()
      (skip): Sequential(
        (0): Conv2d(24, 24, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), groups=24, bias=False)
        (1): BatchNorm2d(24, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
        (2): Conv2d(24, 40, kernel_size=(1, 1), stride=(1, 1))
        (3): BatchNorm2d(40, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      )
    )
    (4): Block(
      (conv1): Conv2d(40, 240, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn1): BatchNorm2d(240, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (act1): Hardswish()
      (conv2): Conv2d(240, 240, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2), groups=240, bias=False)
      (bn2): BatchNorm2d(240, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (act2): Hardswish()
      (se): SeModule(
        (se): Sequential(
          (0): AdaptiveAvgPool2d(output_size=1)
          (1): Conv2d(240, 60, kernel_size=(1, 1), stride=(1, 1), bias=False)
          (2): BatchNorm2d(60, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
          (3): ReLU(inplace=True)
          (4): Conv2d(60, 240, kernel_size=(1, 1), stride=(1, 1), bias=False)
          (5): Hardsigmoid()
        )
      )
      (conv3): Conv2d(240, 40, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn3): BatchNorm2d(40, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (act3): Hardswish()
    )
    (5): Block(
      (conv1): Conv2d(40, 240, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn1): BatchNorm2d(240, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (act1): Hardswish()
      (conv2): Conv2d(240, 240, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2), groups=240, bias=False)
      (bn2): BatchNorm2d(240, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (act2): Hardswish()
      (se): SeModule(
        (se): Sequential(
          (0): AdaptiveAvgPool2d(output_size=1)
          (1): Conv2d(240, 60, kernel_size=(1, 1), stride=(1, 1), bias=False)
          (2): BatchNorm2d(60, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
          (3): ReLU(inplace=True)
          (4): Conv2d(60, 240, kernel_size=(1, 1), stride=(1, 1), bias=False)
          (5): Hardsigmoid()
        )
      )
      (conv3): Conv2d(240, 40, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn3): BatchNorm2d(40, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (act3): Hardswish()
    )
    (6): Block(
      (conv1): Conv2d(40, 120, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn1): BatchNorm2d(120, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (act1): Hardswish()
      (conv2): Conv2d(120, 120, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2), groups=120, bias=False)
      (bn2): BatchNorm2d(120, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (act2): Hardswish()
      (se): SeModule(
        (se): Sequential(
          (0): AdaptiveAvgPool2d(output_size=1)
          (1): Conv2d(120, 30, kernel_size=(1, 1), stride=(1, 1), bias=False)
          (2): BatchNorm2d(30, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
          (3): ReLU(inplace=True)
          (4): Conv2d(30, 120, kernel_size=(1, 1), stride=(1, 1), bias=False)
          (5): Hardsigmoid()
        )
      )
      (conv3): Conv2d(120, 48, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn3): BatchNorm2d(48, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (act3): Hardswish()
      (skip): Sequential(
        (0): Conv2d(40, 48, kernel_size=(1, 1), stride=(1, 1), bias=False)
        (1): BatchNorm2d(48, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      )
    )
    (7): Block(
      (conv1): Conv2d(48, 144, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn1): BatchNorm2d(144, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (act1): Hardswish()
      (conv2): Conv2d(144, 144, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2), groups=144, bias=False)
      (bn2): BatchNorm2d(144, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (act2): Hardswish()
      (se): SeModule(
        (se): Sequential(
          (0): AdaptiveAvgPool2d(output_size=1)
          (1): Conv2d(144, 36, kernel_size=(1, 1), stride=(1, 1), bias=False)
          (2): BatchNorm2d(36, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
          (3): ReLU(inplace=True)
          (4): Conv2d(36, 144, kernel_size=(1, 1), stride=(1, 1), bias=False)
          (5): Hardsigmoid()
        )
      )
      (conv3): Conv2d(144, 48, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn3): BatchNorm2d(48, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (act3): Hardswish()
    )
    (8): Block(
      (conv1): Conv2d(48, 288, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn1): BatchNorm2d(288, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (act1): Hardswish()
      (conv2): Conv2d(288, 288, kernel_size=(5, 5), stride=(2, 2), padding=(2, 2), groups=288, bias=False)
      (bn2): BatchNorm2d(288, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (act2): Hardswish()
      (se): SeModule(
        (se): Sequential(
          (0): AdaptiveAvgPool2d(output_size=1)
          (1): Conv2d(288, 72, kernel_size=(1, 1), stride=(1, 1), bias=False)
          (2): BatchNorm2d(72, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
          (3): ReLU(inplace=True)
          (4): Conv2d(72, 288, kernel_size=(1, 1), stride=(1, 1), bias=False)
          (5): Hardsigmoid()
        )
      )
      (conv3): Conv2d(288, 96, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn3): BatchNorm2d(96, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (act3): Hardswish()
      (skip): Sequential(
        (0): Conv2d(48, 48, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), groups=48, bias=False)
        (1): BatchNorm2d(48, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
        (2): Conv2d(48, 96, kernel_size=(1, 1), stride=(1, 1))
        (3): BatchNorm2d(96, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      )
    )
    (9): Block(
      (conv1): Conv2d(96, 576, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn1): BatchNorm2d(576, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (act1): Hardswish()
      (conv2): Conv2d(576, 576, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2), groups=576, bias=False)
      (bn2): BatchNorm2d(576, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (act2): Hardswish()
      (se): SeModule(
        (se): Sequential(
          (0): AdaptiveAvgPool2d(output_size=1)
          (1): Conv2d(576, 144, kernel_size=(1, 1), stride=(1, 1), bias=False)
          (2): BatchNorm2d(144, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
          (3): ReLU(inplace=True)
          (4): Conv2d(144, 576, kernel_size=(1, 1), stride=(1, 1), bias=False)
          (5): Hardsigmoid()
        )
      )
      (conv3): Conv2d(576, 96, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn3): BatchNorm2d(96, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (act3): Hardswish()
    )
    (10): Block(
      (conv1): Conv2d(96, 576, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn1): BatchNorm2d(576, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (act1): Hardswish()
      (conv2): Conv2d(576, 576, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2), groups=576, bias=False)
      (bn2): BatchNorm2d(576, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (act2): Hardswish()
      (se): SeModule(
        (se): Sequential(
          (0): AdaptiveAvgPool2d(output_size=1)
          (1): Conv2d(576, 144, kernel_size=(1, 1), stride=(1, 1), bias=False)
          (2): BatchNorm2d(144, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
          (3): ReLU(inplace=True)
          (4): Conv2d(144, 576, kernel_size=(1, 1), stride=(1, 1), bias=False)
          (5): Hardsigmoid()
        )
      )
      (conv3): Conv2d(576, 96, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn3): BatchNorm2d(96, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (act3): Hardswish()
    )
  )
  (conv2): Conv2d(96, 576, kernel_size=(1, 1), stride=(1, 1), bias=False)
  (bn2): BatchNorm2d(576, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
  (hs2): Hardswish()
  (gap): AdaptiveAvgPool2d(output_size=1)
  (linear3): Linear(in_features=576, out_features=1280, bias=False)
  (bn3): BatchNorm1d(1280, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
  (hs3): Hardswish()
  (drop): Dropout(p=0.2, inplace=False)
  (linear4): Linear(in_features=1280, out_features=1000, bias=True)
)
number of params: 2950524
LR = 0.00400000
Batch size = 4096
Update frequent = 2
Number of training examples = 1281167
Number of training training per epoch = 312
Param groups = {
  "decay": {
    "weight_decay": 0.05,
    "params": [
      "conv1.weight",
      "bneck.0.conv1.weight",
      "bneck.0.conv2.weight",
      "bneck.0.se.se.1.weight",
      "bneck.0.se.se.4.weight",
      "bneck.0.conv3.weight",
      "bneck.0.skip.0.weight",
      "bneck.1.conv1.weight",
      "bneck.1.conv2.weight",
      "bneck.1.conv3.weight",
      "bneck.1.skip.0.weight",
      "bneck.1.skip.2.weight",
      "bneck.2.conv1.weight",
      "bneck.2.conv2.weight",
      "bneck.2.conv3.weight",
      "bneck.3.conv1.weight",
      "bneck.3.conv2.weight",
      "bneck.3.se.se.1.weight",
      "bneck.3.se.se.4.weight",
      "bneck.3.conv3.weight",
      "bneck.3.skip.0.weight",
      "bneck.3.skip.2.weight",
      "bneck.4.conv1.weight",
      "bneck.4.conv2.weight",
      "bneck.4.se.se.1.weight",
      "bneck.4.se.se.4.weight",
      "bneck.4.conv3.weight",
      "bneck.5.conv1.weight",
      "bneck.5.conv2.weight",
      "bneck.5.se.se.1.weight",
      "bneck.5.se.se.4.weight",
      "bneck.5.conv3.weight",
      "bneck.6.conv1.weight",
      "bneck.6.conv2.weight",
      "bneck.6.se.se.1.weight",
      "bneck.6.se.se.4.weight",
      "bneck.6.conv3.weight",
      "bneck.6.skip.0.weight",
      "bneck.7.conv1.weight",
      "bneck.7.conv2.weight",
      "bneck.7.se.se.1.weight",
      "bneck.7.se.se.4.weight",
      "bneck.7.conv3.weight",
      "bneck.8.conv1.weight",
      "bneck.8.conv2.weight",
      "bneck.8.se.se.1.weight",
      "bneck.8.se.se.4.weight",
      "bneck.8.conv3.weight",
      "bneck.8.skip.0.weight",
      "bneck.8.skip.2.weight",
      "bneck.9.conv1.weight",
      "bneck.9.conv2.weight",
      "bneck.9.se.se.1.weight",
      "bneck.9.se.se.4.weight",
      "bneck.9.conv3.weight",
      "bneck.10.conv1.weight",
      "bneck.10.conv2.weight",
      "bneck.10.se.se.1.weight",
      "bneck.10.se.se.4.weight",
      "bneck.10.conv3.weight",
      "conv2.weight",
      "linear3.weight",
      "linear4.weight"
    ],
    "lr_scale": 1.0
  },
  "no_decay": {
    "weight_decay": 0.0,
    "params": [
      "bn1.weight",
      "bn1.bias",
      "bneck.0.bn1.weight",
      "bneck.0.bn1.bias",
      "bneck.0.bn2.weight",
      "bneck.0.bn2.bias",
      "bneck.0.se.se.2.weight",
      "bneck.0.se.se.2.bias",
      "bneck.0.bn3.weight",
      "bneck.0.bn3.bias",
      "bneck.0.skip.1.weight",
      "bneck.0.skip.1.bias",
      "bneck.1.bn1.weight",
      "bneck.1.bn1.bias",
      "bneck.1.bn2.weight",
      "bneck.1.bn2.bias",
      "bneck.1.bn3.weight",
      "bneck.1.bn3.bias",
      "bneck.1.skip.1.weight",
      "bneck.1.skip.1.bias",
      "bneck.1.skip.2.bias",
      "bneck.1.skip.3.weight",
      "bneck.1.skip.3.bias",
      "bneck.2.bn1.weight",
      "bneck.2.bn1.bias",
      "bneck.2.bn2.weight",
      "bneck.2.bn2.bias",
      "bneck.2.bn3.weight",
      "bneck.2.bn3.bias",
      "bneck.3.bn1.weight",
      "bneck.3.bn1.bias",
      "bneck.3.bn2.weight",
      "bneck.3.bn2.bias",
      "bneck.3.se.se.2.weight",
      "bneck.3.se.se.2.bias",
      "bneck.3.bn3.weight",
      "bneck.3.bn3.bias",
      "bneck.3.skip.1.weight",
      "bneck.3.skip.1.bias",
      "bneck.3.skip.2.bias",
      "bneck.3.skip.3.weight",
      "bneck.3.skip.3.bias",
      "bneck.4.bn1.weight",
      "bneck.4.bn1.bias",
      "bneck.4.bn2.weight",
      "bneck.4.bn2.bias",
      "bneck.4.se.se.2.weight",
      "bneck.4.se.se.2.bias",
      "bneck.4.bn3.weight",
      "bneck.4.bn3.bias",
      "bneck.5.bn1.weight",
      "bneck.5.bn1.bias",
      "bneck.5.bn2.weight",
      "bneck.5.bn2.bias",
      "bneck.5.se.se.2.weight",
      "bneck.5.se.se.2.bias",
      "bneck.5.bn3.weight",
      "bneck.5.bn3.bias",
      "bneck.6.bn1.weight",
      "bneck.6.bn1.bias",
      "bneck.6.bn2.weight",
      "bneck.6.bn2.bias",
      "bneck.6.se.se.2.weight",
      "bneck.6.se.se.2.bias",
      "bneck.6.bn3.weight",
      "bneck.6.bn3.bias",
      "bneck.6.skip.1.weight",
      "bneck.6.skip.1.bias",
      "bneck.7.bn1.weight",
      "bneck.7.bn1.bias",
      "bneck.7.bn2.weight",
      "bneck.7.bn2.bias",
      "bneck.7.se.se.2.weight",
      "bneck.7.se.se.2.bias",
      "bneck.7.bn3.weight",
      "bneck.7.bn3.bias",
      "bneck.8.bn1.weight",
      "bneck.8.bn1.bias",
      "bneck.8.bn2.weight",
      "bneck.8.bn2.bias",
      "bneck.8.se.se.2.weight",
      "bneck.8.se.se.2.bias",
      "bneck.8.bn3.weight",
      "bneck.8.bn3.bias",
      "bneck.8.skip.1.weight",
      "bneck.8.skip.1.bias",
      "bneck.8.skip.2.bias",
      "bneck.8.skip.3.weight",
      "bneck.8.skip.3.bias",
      "bneck.9.bn1.weight",
      "bneck.9.bn1.bias",
      "bneck.9.bn2.weight",
      "bneck.9.bn2.bias",
      "bneck.9.se.se.2.weight",
      "bneck.9.se.se.2.bias",
      "bneck.9.bn3.weight",
      "bneck.9.bn3.bias",
      "bneck.10.bn1.weight",
      "bneck.10.bn1.bias",
      "bneck.10.bn2.weight",
      "bneck.10.bn2.bias",
      "bneck.10.se.se.2.weight",
      "bneck.10.se.se.2.bias",
      "bneck.10.bn3.weight",
      "bneck.10.bn3.bias",
      "bn2.weight",
      "bn2.bias",
      "bn3.weight",
      "bn3.bias",
      "linear4.bias"
    ],
    "lr_scale": 1.0
  }
}
Use Cosine LR scheduler
Set warmup steps = 6240
Set warmup steps = 0
Max WD = 0.0500000, Min WD = 0.0500000
criterion = LabelSmoothingCrossEntropy()
Auto resume checkpoint: 
Start training for 300 epochs
Epoch: [0]  [  0/625]  eta: 5:22:53  lr: 0.000000  min_lr: 0.000000  loss: 6.9073 (6.9073)  class_acc: 0.0000 (0.0000)  weight_decay: 0.0500 (0.0500)  time: 30.9969  data: 24.4808  max mem: 2905
Epoch: [0]  [200/625]  eta: 0:15:28  lr: 0.000064  min_lr: 0.000064  loss: 6.8783 (6.8985)  class_acc: 0.0000 (0.0013)  weight_decay: 0.0500 (0.0500)  grad_norm: 0.4079 (0.4513)  time: 1.9811  data: 0.3600  max mem: 2905
Epoch: [0]  [400/625]  eta: 0:08:01  lr: 0.000128  min_lr: 0.000128  loss: 6.7819 (6.8701)  class_acc: 0.0039 (0.0018)  weight_decay: 0.0500 (0.0500)  grad_norm: 0.5901 (0.4735)  time: 2.2324  data: 0.0009  max mem: 2905
Epoch: [0]  [600/625]  eta: 0:00:53  lr: 0.000192  min_lr: 0.000192  loss: 6.5924 (6.8041)  class_acc: 0.0039 (0.0032)  weight_decay: 0.0500 (0.0500)  grad_norm: 0.9315 (0.5781)  time: 2.0935  data: 0.0452  max mem: 2905
Epoch: [0]  [624/625]  eta: 0:00:02  lr: 0.000199  min_lr: 0.000199  loss: 6.5503 (6.7951)  class_acc: 0.0078 (0.0034)  weight_decay: 0.0500 (0.0500)  grad_norm: 0.9210 (0.5922)  time: 0.6318  data: 0.0064  max mem: 2905
Epoch: [0] Total time: 0:21:52 (2.1002 s / it)
Averaged stats: lr: 0.000199  min_lr: 0.000199  loss: 6.5503 (6.7943)  class_acc: 0.0078 (0.0036)  weight_decay: 0.0500 (0.0500)  grad_norm: 0.9210 (0.5922)
Test:  [ 0/50]  eta: 0:11:29  loss: 6.2122 (6.2122)  acc1: 0.0000 (0.0000)  acc5: 2.4000 (2.4000)  time: 13.7935  data: 13.0781  max mem: 2905
Test:  [10/50]  eta: 0:01:21  loss: 6.2798 (6.2349)  acc1: 0.8000 (2.0364)  acc5: 5.6000 (6.4727)  time: 2.0441  data: 1.9611  max mem: 2905
Test:  [20/50]  eta: 0:00:46  loss: 6.2798 (6.2465)  acc1: 1.6000 (1.9048)  acc5: 5.6000 (6.4381)  time: 0.9553  data: 0.9362  max mem: 2905
Test:  [30/50]  eta: 0:00:27  loss: 6.2106 (6.2254)  acc1: 1.6000 (2.1161)  acc5: 6.4000 (6.7097)  time: 1.0293  data: 1.0112  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 6.2730 (6.2426)  acc1: 1.6000 (1.9707)  acc5: 5.6000 (6.1463)  time: 0.9834  data: 0.9653  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 6.2752 (6.2454)  acc1: 0.8000 (1.8560)  acc5: 4.8000 (6.0320)  time: 0.9657  data: 0.9474  max mem: 2905
Test: Total time: 0:00:59 (1.1806 s / it)
* Acc@1 1.620 Acc@5 5.984 loss 6.224
Accuracy of the model on the 50000 test images: 1.6%
Max accuracy: 1.62%
Epoch: [1]  [  0/625]  eta: 3:30:46  lr: 0.000200  min_lr: 0.000200  loss: 6.5212 (6.5212)  class_acc: 0.0078 (0.0078)  weight_decay: 0.0500 (0.0500)  time: 20.2340  data: 18.8082  max mem: 2905
Epoch: [1]  [200/625]  eta: 0:14:09  lr: 0.000264  min_lr: 0.000264  loss: 6.3831 (6.4747)  class_acc: 0.0156 (0.0131)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.1870 (1.0887)  time: 1.9414  data: 0.0293  max mem: 2905
Epoch: [1]  [400/625]  eta: 0:07:25  lr: 0.000328  min_lr: 0.000328  loss: 6.3248 (6.4075)  class_acc: 0.0234 (0.0163)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.2754 (1.1543)  time: 1.9028  data: 0.0008  max mem: 2905
Epoch: [1]  [600/625]  eta: 0:00:49  lr: 0.000392  min_lr: 0.000392  loss: 6.1402 (6.3414)  class_acc: 0.0273 (0.0195)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.3534 (1.2217)  time: 2.0527  data: 0.0007  max mem: 2905
Epoch: [1]  [624/625]  eta: 0:00:01  lr: 0.000399  min_lr: 0.000399  loss: 6.1052 (6.3331)  class_acc: 0.0312 (0.0201)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.4380 (1.2335)  time: 0.8558  data: 0.0018  max mem: 2905
Epoch: [1] Total time: 0:20:10 (1.9367 s / it)
Averaged stats: lr: 0.000399  min_lr: 0.000399  loss: 6.1052 (6.3370)  class_acc: 0.0312 (0.0199)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.4380 (1.2335)
Test:  [ 0/50]  eta: 0:10:19  loss: 5.3883 (5.3883)  acc1: 5.6000 (5.6000)  acc5: 20.8000 (20.8000)  time: 12.3925  data: 12.3576  max mem: 2905
Test:  [10/50]  eta: 0:01:24  loss: 5.4225 (5.4007)  acc1: 6.4000 (7.0545)  acc5: 20.0000 (18.8364)  time: 2.1014  data: 2.0793  max mem: 2905
Test:  [20/50]  eta: 0:00:51  loss: 5.4212 (5.4345)  acc1: 5.6000 (5.9048)  acc5: 15.2000 (17.8286)  time: 1.1690  data: 1.1492  max mem: 2905
Test:  [30/50]  eta: 0:00:30  loss: 5.3942 (5.4341)  acc1: 4.8000 (5.8839)  acc5: 15.2000 (17.3161)  time: 1.2309  data: 1.2122  max mem: 2905
Test:  [40/50]  eta: 0:00:13  loss: 5.4350 (5.4472)  acc1: 4.8000 (5.6976)  acc5: 14.4000 (16.6049)  time: 0.8849  data: 0.8655  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 5.5120 (5.4625)  acc1: 4.8000 (5.6800)  acc5: 14.4000 (16.3680)  time: 0.8166  data: 0.7956  max mem: 2905
Test: Total time: 0:00:54 (1.0842 s / it)
* Acc@1 6.006 Acc@5 17.070 loss 5.431
Accuracy of the model on the 50000 test images: 6.0%
Max accuracy: 6.01%
Epoch: [2]  [  0/625]  eta: 3:47:03  lr: 0.000400  min_lr: 0.000400  loss: 6.1433 (6.1433)  class_acc: 0.0234 (0.0234)  weight_decay: 0.0500 (0.0500)  time: 21.7982  data: 18.5213  max mem: 2905
Epoch: [2]  [200/625]  eta: 0:13:53  lr: 0.000464  min_lr: 0.000464  loss: 6.0624 (6.0685)  class_acc: 0.0391 (0.0372)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.6316 (1.5256)  time: 1.8131  data: 0.0244  max mem: 2905
Epoch: [2]  [400/625]  eta: 0:07:09  lr: 0.000528  min_lr: 0.000528  loss: 5.8892 (6.0085)  class_acc: 0.0508 (0.0422)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.5450 (1.5541)  time: 1.8028  data: 0.0042  max mem: 2905
Epoch: [2]  [600/625]  eta: 0:00:47  lr: 0.000592  min_lr: 0.000592  loss: 5.7767 (5.9473)  class_acc: 0.0586 (0.0475)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7937 (1.6152)  time: 1.9841  data: 0.0006  max mem: 2905
Epoch: [2]  [624/625]  eta: 0:00:01  lr: 0.000599  min_lr: 0.000599  loss: 5.7209 (5.9391)  class_acc: 0.0625 (0.0482)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7196 (1.6232)  time: 0.8288  data: 0.0015  max mem: 2905
Epoch: [2] Total time: 0:19:34 (1.8798 s / it)
Averaged stats: lr: 0.000599  min_lr: 0.000599  loss: 5.7209 (5.9395)  class_acc: 0.0625 (0.0480)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7196 (1.6232)
Test:  [ 0/50]  eta: 0:09:46  loss: 4.8598 (4.8598)  acc1: 11.2000 (11.2000)  acc5: 27.2000 (27.2000)  time: 11.7263  data: 11.6884  max mem: 2905
Test:  [10/50]  eta: 0:01:17  loss: 4.8561 (4.7968)  acc1: 12.0000 (12.5818)  acc5: 28.8000 (29.4545)  time: 1.9294  data: 1.9089  max mem: 2905
Test:  [20/50]  eta: 0:00:44  loss: 4.8561 (4.8427)  acc1: 11.2000 (11.2762)  acc5: 28.8000 (28.1524)  time: 0.9804  data: 0.9605  max mem: 2905
Test:  [30/50]  eta: 0:00:25  loss: 4.8037 (4.8281)  acc1: 10.4000 (11.1742)  acc5: 28.0000 (28.5419)  time: 0.9209  data: 0.9003  max mem: 2905
Test:  [40/50]  eta: 0:00:10  loss: 4.8505 (4.8551)  acc1: 8.8000 (10.9659)  acc5: 27.2000 (28.1756)  time: 0.6264  data: 0.6058  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 4.9349 (4.8786)  acc1: 9.6000 (10.8160)  acc5: 24.8000 (27.5040)  time: 0.6236  data: 0.6034  max mem: 2905
Test: Total time: 0:00:46 (0.9271 s / it)
* Acc@1 11.498 Acc@5 28.242 loss 4.837
Accuracy of the model on the 50000 test images: 11.5%
Max accuracy: 11.50%
Epoch: [3]  [  0/625]  eta: 3:34:40  lr: 0.000600  min_lr: 0.000600  loss: 5.7872 (5.7872)  class_acc: 0.0508 (0.0508)  weight_decay: 0.0500 (0.0500)  time: 20.6085  data: 20.4796  max mem: 2905
Epoch: [3]  [200/625]  eta: 0:13:29  lr: 0.000664  min_lr: 0.000664  loss: 5.6046 (5.6750)  class_acc: 0.0742 (0.0743)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9255 (1.8777)  time: 1.8261  data: 0.2726  max mem: 2905
Epoch: [3]  [400/625]  eta: 0:07:08  lr: 0.000728  min_lr: 0.000728  loss: 5.4507 (5.6158)  class_acc: 0.1055 (0.0807)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1282 (1.9889)  time: 1.8797  data: 0.0230  max mem: 2905
Epoch: [3]  [600/625]  eta: 0:00:48  lr: 0.000792  min_lr: 0.000792  loss: 5.4681 (5.5621)  class_acc: 0.1016 (0.0856)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9530 (2.0324)  time: 1.9466  data: 0.0006  max mem: 2905
Epoch: [3]  [624/625]  eta: 0:00:01  lr: 0.000799  min_lr: 0.000799  loss: 5.3850 (5.5557)  class_acc: 0.1055 (0.0865)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1287 (2.0300)  time: 0.7554  data: 0.0016  max mem: 2905
Epoch: [3] Total time: 0:19:31 (1.8738 s / it)
Averaged stats: lr: 0.000799  min_lr: 0.000799  loss: 5.3850 (5.5575)  class_acc: 0.1055 (0.0864)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1287 (2.0300)
Test:  [ 0/50]  eta: 0:10:33  loss: 4.4218 (4.4218)  acc1: 12.8000 (12.8000)  acc5: 36.0000 (36.0000)  time: 12.6686  data: 12.6427  max mem: 2905
Test:  [10/50]  eta: 0:01:15  loss: 4.3667 (4.2635)  acc1: 19.2000 (18.6182)  acc5: 40.0000 (40.8727)  time: 1.8861  data: 1.8650  max mem: 2905
Test:  [20/50]  eta: 0:00:40  loss: 4.3667 (4.3161)  acc1: 16.0000 (16.3810)  acc5: 39.2000 (38.1333)  time: 0.7907  data: 0.7708  max mem: 2905
Test:  [30/50]  eta: 0:00:25  loss: 4.3468 (4.2962)  acc1: 15.2000 (16.4903)  acc5: 36.8000 (38.0903)  time: 0.9523  data: 0.9335  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 4.2393 (4.3090)  acc1: 16.8000 (16.9366)  acc5: 36.8000 (37.7951)  time: 1.0633  data: 1.0448  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 4.4039 (4.3490)  acc1: 16.0000 (16.4960)  acc5: 34.4000 (37.3280)  time: 0.7165  data: 0.6985  max mem: 2905
Test: Total time: 0:00:54 (1.0855 s / it)
* Acc@1 17.338 Acc@5 37.778 loss 4.303
Accuracy of the model on the 50000 test images: 17.3%
Max accuracy: 17.34%
Epoch: [4]  [  0/625]  eta: 3:15:07  lr: 0.000800  min_lr: 0.000800  loss: 5.2683 (5.2683)  class_acc: 0.1016 (0.1016)  weight_decay: 0.0500 (0.0500)  time: 18.7315  data: 18.1657  max mem: 2905
Epoch: [4]  [200/625]  eta: 0:13:45  lr: 0.000864  min_lr: 0.000864  loss: 5.3035 (5.3295)  class_acc: 0.1133 (0.1126)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0196 (2.1965)  time: 1.8294  data: 0.0499  max mem: 2905
Epoch: [4]  [400/625]  eta: 0:07:13  lr: 0.000928  min_lr: 0.000928  loss: 5.2115 (5.2869)  class_acc: 0.1328 (0.1181)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1358 (inf)  time: 1.9779  data: 0.0099  max mem: 2905
Epoch: [4]  [600/625]  eta: 0:00:48  lr: 0.000992  min_lr: 0.000992  loss: 5.1308 (5.2411)  class_acc: 0.1406 (0.1237)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1777 (inf)  time: 1.9683  data: 0.0520  max mem: 2905
Epoch: [4]  [624/625]  eta: 0:00:01  lr: 0.001000  min_lr: 0.001000  loss: 5.1033 (5.2360)  class_acc: 0.1484 (0.1244)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3770 (inf)  time: 0.8879  data: 0.0299  max mem: 2905
Epoch: [4] Total time: 0:19:35 (1.8809 s / it)
Averaged stats: lr: 0.001000  min_lr: 0.001000  loss: 5.1033 (5.2305)  class_acc: 0.1484 (0.1258)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3770 (inf)
Test:  [ 0/50]  eta: 0:10:10  loss: 3.8296 (3.8296)  acc1: 24.8000 (24.8000)  acc5: 44.0000 (44.0000)  time: 12.2025  data: 12.1732  max mem: 2905
Test:  [10/50]  eta: 0:01:29  loss: 3.8296 (3.8120)  acc1: 24.8000 (25.7455)  acc5: 48.0000 (47.7091)  time: 2.2404  data: 2.2214  max mem: 2905
Test:  [20/50]  eta: 0:00:52  loss: 3.9097 (3.9069)  acc1: 21.6000 (22.8571)  acc5: 46.4000 (45.1810)  time: 1.2378  data: 1.2191  max mem: 2905
Test:  [30/50]  eta: 0:00:29  loss: 3.9239 (3.8921)  acc1: 20.8000 (23.0968)  acc5: 41.6000 (45.0839)  time: 1.0447  data: 1.0252  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 3.9548 (3.9210)  acc1: 22.4000 (22.8683)  acc5: 41.6000 (44.7610)  time: 0.6728  data: 0.6523  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 4.0196 (3.9464)  acc1: 20.8000 (22.3360)  acc5: 42.4000 (44.3680)  time: 0.6484  data: 0.6262  max mem: 2905
Test: Total time: 0:00:52 (1.0593 s / it)
* Acc@1 22.582 Acc@5 44.934 loss 3.915
Accuracy of the model on the 50000 test images: 22.6%
Max accuracy: 22.58%
Epoch: [5]  [  0/625]  eta: 3:37:49  lr: 0.001000  min_lr: 0.001000  loss: 5.1298 (5.1298)  class_acc: 0.1367 (0.1367)  weight_decay: 0.0500 (0.0500)  time: 20.9108  data: 18.1586  max mem: 2905
Epoch: [5]  [200/625]  eta: 0:13:51  lr: 0.001064  min_lr: 0.001064  loss: 5.0483 (5.0432)  class_acc: 0.1602 (0.1535)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.8175 (2.6947)  time: 1.9335  data: 0.0055  max mem: 2905
Epoch: [5]  [400/625]  eta: 0:07:11  lr: 0.001128  min_lr: 0.001128  loss: 4.9352 (5.0098)  class_acc: 0.1641 (0.1558)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1319 (2.5978)  time: 2.0246  data: 0.0043  max mem: 2905
Epoch: [5]  [600/625]  eta: 0:00:48  lr: 0.001192  min_lr: 0.001192  loss: 4.8698 (4.9720)  class_acc: 0.1641 (0.1608)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.8755 (2.6583)  time: 1.9903  data: 0.0335  max mem: 2905
Epoch: [5]  [624/625]  eta: 0:00:01  lr: 0.001200  min_lr: 0.001200  loss: 4.8149 (4.9663)  class_acc: 0.1719 (0.1614)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6388 (2.6477)  time: 0.7885  data: 0.0014  max mem: 2905
Epoch: [5] Total time: 0:19:32 (1.8765 s / it)
Averaged stats: lr: 0.001200  min_lr: 0.001200  loss: 4.8149 (4.9633)  class_acc: 0.1719 (0.1627)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6388 (2.6477)
Test:  [ 0/50]  eta: 0:10:16  loss: 3.5700 (3.5700)  acc1: 28.8000 (28.8000)  acc5: 50.4000 (50.4000)  time: 12.3370  data: 12.3068  max mem: 2905
Test:  [10/50]  eta: 0:01:23  loss: 3.5700 (3.5258)  acc1: 28.8000 (27.9273)  acc5: 52.0000 (52.3636)  time: 2.0841  data: 2.0649  max mem: 2905
Test:  [20/50]  eta: 0:00:48  loss: 3.6373 (3.6148)  acc1: 25.6000 (25.3714)  acc5: 50.4000 (50.6667)  time: 1.0670  data: 1.0480  max mem: 2905
Test:  [30/50]  eta: 0:00:27  loss: 3.6760 (3.5991)  acc1: 24.8000 (26.0387)  acc5: 49.6000 (50.6839)  time: 0.9670  data: 0.9467  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 3.6760 (3.6394)  acc1: 25.6000 (26.1659)  acc5: 49.6000 (49.9122)  time: 0.6360  data: 0.6151  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 3.6795 (3.6541)  acc1: 25.6000 (25.9040)  acc5: 50.4000 (50.0000)  time: 0.5791  data: 0.5590  max mem: 2905
Test: Total time: 0:00:48 (0.9651 s / it)
* Acc@1 26.876 Acc@5 50.772 loss 3.622
Accuracy of the model on the 50000 test images: 26.9%
Max accuracy: 26.88%
Epoch: [6]  [  0/625]  eta: 3:29:02  lr: 0.001200  min_lr: 0.001200  loss: 4.9009 (4.9009)  class_acc: 0.1836 (0.1836)  weight_decay: 0.0500 (0.0500)  time: 20.0682  data: 19.9396  max mem: 2905
Epoch: [6]  [200/625]  eta: 0:13:53  lr: 0.001264  min_lr: 0.001264  loss: 4.7983 (4.7987)  class_acc: 0.1953 (0.1882)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6991 (2.5532)  time: 1.9000  data: 1.6467  max mem: 2905
Epoch: [6]  [400/625]  eta: 0:07:08  lr: 0.001328  min_lr: 0.001328  loss: 4.7366 (4.7691)  class_acc: 0.1992 (0.1925)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5836 (2.5758)  time: 1.6145  data: 1.4355  max mem: 2905
Epoch: [6]  [600/625]  eta: 0:00:47  lr: 0.001393  min_lr: 0.001393  loss: 4.6280 (4.7387)  class_acc: 0.2070 (0.1974)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2042 (2.5552)  time: 2.0588  data: 1.8792  max mem: 2905
Epoch: [6]  [624/625]  eta: 0:00:01  lr: 0.001400  min_lr: 0.001400  loss: 4.6438 (4.7372)  class_acc: 0.2070 (0.1974)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6381 (2.5659)  time: 0.6559  data: 0.4986  max mem: 2905
Epoch: [6] Total time: 0:19:26 (1.8664 s / it)
Averaged stats: lr: 0.001400  min_lr: 0.001400  loss: 4.6438 (4.7416)  class_acc: 0.2070 (0.1962)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6381 (2.5659)
Test:  [ 0/50]  eta: 0:09:43  loss: 3.2068 (3.2068)  acc1: 33.6000 (33.6000)  acc5: 59.2000 (59.2000)  time: 11.6622  data: 11.6334  max mem: 2905
Test:  [10/50]  eta: 0:01:24  loss: 3.2586 (3.2402)  acc1: 34.4000 (33.4545)  acc5: 60.0000 (59.3455)  time: 2.1058  data: 2.0845  max mem: 2905
Test:  [20/50]  eta: 0:00:49  loss: 3.3495 (3.3518)  acc1: 30.4000 (30.8571)  acc5: 57.6000 (56.6476)  time: 1.1618  data: 1.1412  max mem: 2905
Test:  [30/50]  eta: 0:00:26  loss: 3.3963 (3.3467)  acc1: 29.6000 (30.6065)  acc5: 54.4000 (56.4903)  time: 0.9146  data: 0.8941  max mem: 2905
Test:  [40/50]  eta: 0:00:10  loss: 3.3060 (3.3673)  acc1: 29.6000 (30.2829)  acc5: 55.2000 (55.8829)  time: 0.4684  data: 0.4486  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 3.4843 (3.3869)  acc1: 28.8000 (29.8400)  acc5: 53.6000 (55.4240)  time: 0.4594  data: 0.4407  max mem: 2905
Test: Total time: 0:00:45 (0.9178 s / it)
* Acc@1 30.902 Acc@5 55.788 loss 3.356
Accuracy of the model on the 50000 test images: 30.9%
Max accuracy: 30.90%
Epoch: [7]  [  0/625]  eta: 4:03:25  lr: 0.001400  min_lr: 0.001400  loss: 4.6585 (4.6585)  class_acc: 0.2344 (0.2344)  weight_decay: 0.0500 (0.0500)  time: 23.3680  data: 16.2731  max mem: 2905
Epoch: [7]  [200/625]  eta: 0:13:40  lr: 0.001464  min_lr: 0.001464  loss: 4.5794 (4.6002)  class_acc: 0.2070 (0.2162)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3339 (2.6517)  time: 2.0130  data: 0.0005  max mem: 2905
Epoch: [7]  [400/625]  eta: 0:07:09  lr: 0.001528  min_lr: 0.001528  loss: 4.5334 (4.5822)  class_acc: 0.2305 (0.2191)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3935 (2.6114)  time: 1.9204  data: 0.0005  max mem: 2905
Epoch: [7]  [600/625]  eta: 0:00:47  lr: 0.001593  min_lr: 0.001593  loss: 4.4742 (4.5610)  class_acc: 0.2422 (0.2231)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2934 (2.5822)  time: 2.0638  data: 0.0005  max mem: 2905
Epoch: [7]  [624/625]  eta: 0:00:01  lr: 0.001600  min_lr: 0.001600  loss: 4.4444 (4.5575)  class_acc: 0.2383 (0.2238)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4272 (2.5888)  time: 0.7204  data: 0.0013  max mem: 2905
Epoch: [7] Total time: 0:19:30 (1.8731 s / it)
Averaged stats: lr: 0.001600  min_lr: 0.001600  loss: 4.4444 (4.5618)  class_acc: 0.2383 (0.2250)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4272 (2.5888)
Test:  [ 0/50]  eta: 0:10:50  loss: 2.9093 (2.9093)  acc1: 36.8000 (36.8000)  acc5: 63.2000 (63.2000)  time: 13.0051  data: 12.9644  max mem: 2905
Test:  [10/50]  eta: 0:01:27  loss: 2.9527 (3.0195)  acc1: 37.6000 (37.1636)  acc5: 62.4000 (62.3273)  time: 2.1962  data: 2.1751  max mem: 2905
Test:  [20/50]  eta: 0:00:51  loss: 3.2389 (3.1744)  acc1: 35.2000 (34.0952)  acc5: 58.4000 (59.2762)  time: 1.1524  data: 1.1297  max mem: 2905
Test:  [30/50]  eta: 0:00:29  loss: 3.3120 (3.1727)  acc1: 31.2000 (34.0129)  acc5: 56.8000 (59.1226)  time: 1.0619  data: 1.0387  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 3.2125 (3.1923)  acc1: 32.0000 (34.0098)  acc5: 57.6000 (58.9463)  time: 0.6523  data: 0.6316  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 3.3117 (3.2221)  acc1: 32.0000 (33.6320)  acc5: 57.6000 (58.6240)  time: 0.6373  data: 0.6169  max mem: 2905
Test: Total time: 0:00:49 (0.9976 s / it)
* Acc@1 34.238 Acc@5 59.060 loss 3.189
Accuracy of the model on the 50000 test images: 34.2%
Max accuracy: 34.24%
Epoch: [8]  [  0/625]  eta: 3:37:37  lr: 0.001600  min_lr: 0.001600  loss: 4.4698 (4.4698)  class_acc: 0.2773 (0.2773)  weight_decay: 0.0500 (0.0500)  time: 20.8919  data: 17.4404  max mem: 2905
Epoch: [8]  [200/625]  eta: 0:13:38  lr: 0.001664  min_lr: 0.001664  loss: 4.4204 (4.4612)  class_acc: 0.2422 (0.2399)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5189 (2.6102)  time: 1.8606  data: 0.0007  max mem: 2905
Epoch: [8]  [400/625]  eta: 0:07:10  lr: 0.001728  min_lr: 0.001728  loss: 4.4202 (4.4448)  class_acc: 0.2461 (0.2436)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3623 (2.5957)  time: 1.9448  data: 0.0006  max mem: 2905
Epoch: [8]  [600/625]  eta: 0:00:48  lr: 0.001793  min_lr: 0.001793  loss: 4.3859 (4.4276)  class_acc: 0.2617 (0.2474)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4784 (2.5829)  time: 2.0012  data: 0.0006  max mem: 2905
Epoch: [8]  [624/625]  eta: 0:00:01  lr: 0.001800  min_lr: 0.001800  loss: 4.3258 (4.4244)  class_acc: 0.2656 (0.2481)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2915 (2.5702)  time: 0.7980  data: 0.0013  max mem: 2905
Epoch: [8] Total time: 0:19:42 (1.8919 s / it)
Averaged stats: lr: 0.001800  min_lr: 0.001800  loss: 4.3258 (4.4180)  class_acc: 0.2656 (0.2493)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2915 (2.5702)
Test:  [ 0/50]  eta: 0:09:58  loss: 2.8935 (2.8935)  acc1: 39.2000 (39.2000)  acc5: 61.6000 (61.6000)  time: 11.9733  data: 11.9122  max mem: 2905
Test:  [10/50]  eta: 0:01:29  loss: 2.8935 (2.9067)  acc1: 40.0000 (40.0727)  acc5: 67.2000 (64.8727)  time: 2.2268  data: 2.2035  max mem: 2905
Test:  [20/50]  eta: 0:00:54  loss: 3.0836 (3.0588)  acc1: 36.0000 (36.6095)  acc5: 61.6000 (61.8667)  time: 1.3003  data: 1.2802  max mem: 2905
Test:  [30/50]  eta: 0:00:31  loss: 3.0992 (3.0312)  acc1: 34.4000 (36.9032)  acc5: 60.8000 (62.0387)  time: 1.1798  data: 1.1599  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 3.0874 (3.0463)  acc1: 38.4000 (36.7415)  acc5: 60.8000 (61.8341)  time: 0.7516  data: 0.7325  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 3.0975 (3.0647)  acc1: 35.2000 (36.2720)  acc5: 60.8000 (61.3920)  time: 0.6948  data: 0.6749  max mem: 2905
Test: Total time: 0:00:54 (1.0808 s / it)
* Acc@1 36.742 Acc@5 61.776 loss 3.036
Accuracy of the model on the 50000 test images: 36.7%
Max accuracy: 36.74%
Epoch: [9]  [  0/625]  eta: 3:36:02  lr: 0.001800  min_lr: 0.001800  loss: 4.3272 (4.3272)  class_acc: 0.2500 (0.2500)  weight_decay: 0.0500 (0.0500)  time: 20.7408  data: 16.6444  max mem: 2905
Epoch: [9]  [200/625]  eta: 0:14:23  lr: 0.001864  min_lr: 0.001864  loss: 4.2635 (4.3240)  class_acc: 0.2734 (0.2654)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4228 (2.4175)  time: 1.9545  data: 0.0091  max mem: 2905
Epoch: [9]  [400/625]  eta: 0:07:29  lr: 0.001929  min_lr: 0.001929  loss: 4.2678 (4.3155)  class_acc: 0.2578 (0.2662)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1543 (2.5136)  time: 1.9232  data: 0.0007  max mem: 2905
Epoch: [9]  [600/625]  eta: 0:00:49  lr: 0.001993  min_lr: 0.001993  loss: 4.2901 (4.3029)  class_acc: 0.2617 (0.2689)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5049 (2.5141)  time: 1.9850  data: 0.0731  max mem: 2905
Epoch: [9]  [624/625]  eta: 0:00:01  lr: 0.002000  min_lr: 0.002000  loss: 4.2756 (4.3026)  class_acc: 0.2656 (0.2688)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2845 (2.5103)  time: 0.7254  data: 0.0210  max mem: 2905
Epoch: [9] Total time: 0:20:10 (1.9361 s / it)
Averaged stats: lr: 0.002000  min_lr: 0.002000  loss: 4.2756 (4.2989)  class_acc: 0.2656 (0.2701)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2845 (2.5103)
Test:  [ 0/50]  eta: 0:08:55  loss: 2.7651 (2.7651)  acc1: 43.2000 (43.2000)  acc5: 67.2000 (67.2000)  time: 10.7198  data: 10.6874  max mem: 2905
Test:  [10/50]  eta: 0:01:19  loss: 2.7537 (2.7308)  acc1: 40.8000 (41.3818)  acc5: 67.2000 (66.6909)  time: 1.9963  data: 1.9748  max mem: 2905
Test:  [20/50]  eta: 0:00:48  loss: 2.9111 (2.8878)  acc1: 37.6000 (38.1333)  acc5: 63.2000 (64.1905)  time: 1.1607  data: 1.1401  max mem: 2905
Test:  [30/50]  eta: 0:00:28  loss: 2.9691 (2.8784)  acc1: 36.0000 (38.1935)  acc5: 61.6000 (64.1548)  time: 1.1207  data: 1.1007  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 3.0237 (2.9040)  acc1: 37.6000 (38.1073)  acc5: 61.6000 (63.2781)  time: 0.6884  data: 0.6686  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 3.0316 (2.9240)  acc1: 36.8000 (37.6480)  acc5: 60.8000 (62.9120)  time: 0.5287  data: 0.5094  max mem: 2905
Test: Total time: 0:00:48 (0.9770 s / it)
* Acc@1 38.540 Acc@5 63.834 loss 2.886
Accuracy of the model on the 50000 test images: 38.5%
Max accuracy: 38.54%
Epoch: [10]  [  0/625]  eta: 3:53:22  lr: 0.002000  min_lr: 0.002000  loss: 4.2604 (4.2604)  class_acc: 0.2812 (0.2812)  weight_decay: 0.0500 (0.0500)  time: 22.4036  data: 16.9625  max mem: 2905
Epoch: [10]  [200/625]  eta: 0:14:06  lr: 0.002064  min_lr: 0.002064  loss: 4.2511 (4.2439)  class_acc: 0.2852 (0.2789)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1119 (2.3423)  time: 1.9231  data: 0.0280  max mem: 2905
Epoch: [10]  [400/625]  eta: 0:07:13  lr: 0.002129  min_lr: 0.002129  loss: 4.1512 (4.2185)  class_acc: 0.2969 (0.2835)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5338 (2.3170)  time: 1.8732  data: 0.0006  max mem: 2905
Epoch: [10]  [600/625]  eta: 0:00:48  lr: 0.002193  min_lr: 0.002193  loss: 4.1442 (4.2089)  class_acc: 0.3008 (0.2857)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9480 (inf)  time: 2.0251  data: 0.0221  max mem: 2905
Epoch: [10]  [624/625]  eta: 0:00:01  lr: 0.002200  min_lr: 0.002200  loss: 4.1306 (4.2070)  class_acc: 0.2812 (0.2856)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9993 (inf)  time: 0.7978  data: 0.0014  max mem: 2905
Epoch: [10] Total time: 0:19:35 (1.8811 s / it)
Averaged stats: lr: 0.002200  min_lr: 0.002200  loss: 4.1306 (4.2042)  class_acc: 0.2812 (0.2866)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9993 (inf)
Test:  [ 0/50]  eta: 0:10:14  loss: 2.6962 (2.6962)  acc1: 43.2000 (43.2000)  acc5: 64.8000 (64.8000)  time: 12.2856  data: 12.2574  max mem: 2905
Test:  [10/50]  eta: 0:01:22  loss: 2.6962 (2.6542)  acc1: 46.4000 (44.2909)  acc5: 68.0000 (68.0000)  time: 2.0653  data: 2.0455  max mem: 2905
Test:  [20/50]  eta: 0:00:48  loss: 2.8521 (2.8162)  acc1: 39.2000 (40.6476)  acc5: 65.6000 (66.1333)  time: 1.0829  data: 1.0643  max mem: 2905
Test:  [30/50]  eta: 0:00:28  loss: 2.8902 (2.8010)  acc1: 37.6000 (40.8000)  acc5: 65.6000 (66.2710)  time: 1.0456  data: 1.0258  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 2.8382 (2.8281)  acc1: 38.4000 (40.6829)  acc5: 65.6000 (65.5024)  time: 0.7540  data: 0.7338  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.9067 (2.8491)  acc1: 37.6000 (40.1280)  acc5: 65.6000 (65.2000)  time: 0.7291  data: 0.7101  max mem: 2905
Test: Total time: 0:00:50 (1.0079 s / it)
* Acc@1 40.592 Acc@5 65.836 loss 2.813
Accuracy of the model on the 50000 test images: 40.6%
Max accuracy: 40.59%
Epoch: [11]  [  0/625]  eta: 3:43:47  lr: 0.002200  min_lr: 0.002200  loss: 3.9268 (3.9268)  class_acc: 0.3516 (0.3516)  weight_decay: 0.0500 (0.0500)  time: 21.4832  data: 17.0127  max mem: 2905
Epoch: [11]  [200/625]  eta: 0:14:26  lr: 0.002264  min_lr: 0.002264  loss: 4.1573 (4.1289)  class_acc: 0.3008 (0.3006)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0794 (2.3988)  time: 1.9358  data: 0.0012  max mem: 2905
Epoch: [11]  [400/625]  eta: 0:07:24  lr: 0.002329  min_lr: 0.002329  loss: 4.1037 (4.1273)  class_acc: 0.2930 (0.2998)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4230 (2.4047)  time: 1.8472  data: 0.0245  max mem: 2905
Epoch: [11]  [600/625]  eta: 0:00:49  lr: 0.002393  min_lr: 0.002393  loss: 4.1005 (4.1241)  class_acc: 0.2891 (0.3001)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.6508 (2.3153)  time: 2.1331  data: 0.0204  max mem: 2905
Epoch: [11]  [624/625]  eta: 0:00:01  lr: 0.002400  min_lr: 0.002400  loss: 4.0749 (4.1226)  class_acc: 0.3047 (0.3005)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1353 (2.3125)  time: 0.7291  data: 0.0015  max mem: 2905
Epoch: [11] Total time: 0:20:11 (1.9381 s / it)
Averaged stats: lr: 0.002400  min_lr: 0.002400  loss: 4.0749 (4.1200)  class_acc: 0.3047 (0.3013)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1353 (2.3125)
Test:  [ 0/50]  eta: 0:10:36  loss: 2.4763 (2.4763)  acc1: 46.4000 (46.4000)  acc5: 71.2000 (71.2000)  time: 12.7394  data: 12.7065  max mem: 2905
Test:  [10/50]  eta: 0:01:27  loss: 2.5527 (2.5122)  acc1: 46.4000 (46.8364)  acc5: 71.2000 (70.9091)  time: 2.1897  data: 2.1654  max mem: 2905
Test:  [20/50]  eta: 0:00:53  loss: 2.6607 (2.6901)  acc1: 44.8000 (43.1238)  acc5: 67.2000 (68.1524)  time: 1.2351  data: 1.2141  max mem: 2905
Test:  [30/50]  eta: 0:00:32  loss: 2.7662 (2.6840)  acc1: 40.8000 (42.6581)  acc5: 65.6000 (67.9484)  time: 1.3126  data: 1.2931  max mem: 2905
Test:  [40/50]  eta: 0:00:14  loss: 2.7282 (2.7012)  acc1: 41.6000 (42.7707)  acc5: 68.0000 (67.7659)  time: 1.0129  data: 0.9931  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.7282 (2.7157)  acc1: 40.8000 (42.4960)  acc5: 68.0000 (67.6000)  time: 0.8635  data: 0.8440  max mem: 2905
Test: Total time: 0:00:58 (1.1762 s / it)
* Acc@1 42.952 Acc@5 67.950 loss 2.687
Accuracy of the model on the 50000 test images: 43.0%
Max accuracy: 42.95%
Epoch: [12]  [  0/625]  eta: 3:48:46  lr: 0.002400  min_lr: 0.002400  loss: 4.2393 (4.2393)  class_acc: 0.2773 (0.2773)  weight_decay: 0.0500 (0.0500)  time: 21.9627  data: 18.4014  max mem: 2905
Epoch: [12]  [200/625]  eta: 0:14:12  lr: 0.002464  min_lr: 0.002464  loss: 4.0651 (4.0561)  class_acc: 0.3047 (0.3143)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0893 (2.3006)  time: 1.7633  data: 0.3592  max mem: 2905
Epoch: [12]  [400/625]  eta: 0:07:15  lr: 0.002529  min_lr: 0.002529  loss: 4.0072 (4.0528)  class_acc: 0.3164 (0.3130)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0570 (2.3277)  time: 1.8035  data: 0.0569  max mem: 2905
Epoch: [12]  [600/625]  eta: 0:00:48  lr: 0.002593  min_lr: 0.002593  loss: 4.0482 (4.0528)  class_acc: 0.3203 (0.3136)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9792 (2.3293)  time: 2.1068  data: 0.0625  max mem: 2905
Epoch: [12]  [624/625]  eta: 0:00:01  lr: 0.002600  min_lr: 0.002600  loss: 4.0376 (4.0532)  class_acc: 0.3125 (0.3135)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0785 (2.3239)  time: 0.8143  data: 0.0186  max mem: 2905
Epoch: [12] Total time: 0:20:07 (1.9318 s / it)
Averaged stats: lr: 0.002600  min_lr: 0.002600  loss: 4.0376 (4.0536)  class_acc: 0.3125 (0.3139)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0785 (2.3239)
Test:  [ 0/50]  eta: 0:11:07  loss: 2.4376 (2.4376)  acc1: 46.4000 (46.4000)  acc5: 71.2000 (71.2000)  time: 13.3431  data: 13.3089  max mem: 2905
Test:  [10/50]  eta: 0:01:31  loss: 2.4719 (2.5471)  acc1: 47.2000 (46.0364)  acc5: 70.4000 (69.7455)  time: 2.2773  data: 2.2560  max mem: 2905
Test:  [20/50]  eta: 0:00:55  loss: 2.7218 (2.7209)  acc1: 41.6000 (41.6762)  acc5: 67.2000 (67.6952)  time: 1.2851  data: 1.2659  max mem: 2905
Test:  [30/50]  eta: 0:00:31  loss: 2.8736 (2.7055)  acc1: 39.2000 (41.8581)  acc5: 66.4000 (67.5097)  time: 1.1669  data: 1.1483  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 2.7616 (2.7235)  acc1: 40.8000 (41.5024)  acc5: 67.2000 (67.0829)  time: 0.6777  data: 0.6549  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.8082 (2.7487)  acc1: 40.0000 (41.5040)  acc5: 65.6000 (66.6880)  time: 0.6488  data: 0.6265  max mem: 2905
Test: Total time: 0:00:53 (1.0601 s / it)
* Acc@1 42.014 Acc@5 67.190 loss 2.726
Accuracy of the model on the 50000 test images: 42.0%
Max accuracy: 42.95%
Epoch: [13]  [  0/625]  eta: 4:13:58  lr: 0.002600  min_lr: 0.002600  loss: 3.8410 (3.8410)  class_acc: 0.3633 (0.3633)  weight_decay: 0.0500 (0.0500)  time: 24.3811  data: 23.5717  max mem: 2905
Epoch: [13]  [200/625]  eta: 0:15:25  lr: 0.002665  min_lr: 0.002665  loss: 3.9449 (3.9901)  class_acc: 0.3320 (0.3244)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1871 (2.3120)  time: 1.8859  data: 0.0125  max mem: 2905
Epoch: [13]  [400/625]  eta: 0:07:40  lr: 0.002729  min_lr: 0.002729  loss: 4.0058 (3.9963)  class_acc: 0.3125 (0.3242)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2123 (2.2978)  time: 1.8758  data: 0.0232  max mem: 2905
Epoch: [13]  [600/625]  eta: 0:00:52  lr: 0.002793  min_lr: 0.002793  loss: 3.9816 (3.9945)  class_acc: 0.3125 (0.3250)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0018 (2.2742)  time: 2.2631  data: 0.0006  max mem: 2905
Epoch: [13]  [624/625]  eta: 0:00:02  lr: 0.002800  min_lr: 0.002800  loss: 3.9411 (3.9933)  class_acc: 0.3320 (0.3251)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9408 (2.2684)  time: 1.0124  data: 0.0016  max mem: 2905
Epoch: [13] Total time: 0:21:28 (2.0615 s / it)
Averaged stats: lr: 0.002800  min_lr: 0.002800  loss: 3.9411 (3.9959)  class_acc: 0.3320 (0.3249)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9408 (2.2684)
Test:  [ 0/50]  eta: 0:11:37  loss: 2.5089 (2.5089)  acc1: 45.6000 (45.6000)  acc5: 76.8000 (76.8000)  time: 13.9592  data: 13.9329  max mem: 2905
Test:  [10/50]  eta: 0:01:35  loss: 2.5266 (2.5857)  acc1: 47.2000 (47.0545)  acc5: 70.4000 (70.5455)  time: 2.3968  data: 2.3777  max mem: 2905
Test:  [20/50]  eta: 0:00:54  loss: 2.6529 (2.6943)  acc1: 44.0000 (43.2381)  acc5: 68.8000 (69.2571)  time: 1.2106  data: 1.1920  max mem: 2905
Test:  [30/50]  eta: 0:00:31  loss: 2.6981 (2.6633)  acc1: 40.0000 (43.2774)  acc5: 68.8000 (69.6516)  time: 1.1508  data: 1.1319  max mem: 2905
Test:  [40/50]  eta: 0:00:13  loss: 2.7048 (2.6942)  acc1: 40.8000 (42.6732)  acc5: 68.0000 (68.8195)  time: 0.9040  data: 0.8853  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.7685 (2.7166)  acc1: 40.8000 (42.3680)  acc5: 66.4000 (68.2560)  time: 0.8507  data: 0.8325  max mem: 2905
Test: Total time: 0:01:00 (1.2075 s / it)
* Acc@1 42.820 Acc@5 68.040 loss 2.698
Accuracy of the model on the 50000 test images: 42.8%
Max accuracy: 42.95%
Epoch: [14]  [  0/625]  eta: 4:22:44  lr: 0.002800  min_lr: 0.002800  loss: 3.6967 (3.6967)  class_acc: 0.3594 (0.3594)  weight_decay: 0.0500 (0.0500)  time: 25.2232  data: 25.0372  max mem: 2905
Epoch: [14]  [200/625]  eta: 0:15:03  lr: 0.002865  min_lr: 0.002865  loss: 3.8945 (3.9471)  class_acc: 0.3359 (0.3323)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8518 (2.3200)  time: 1.9372  data: 0.0008  max mem: 2905
Epoch: [14]  [400/625]  eta: 0:07:35  lr: 0.002929  min_lr: 0.002929  loss: 3.9635 (3.9485)  class_acc: 0.3164 (0.3321)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0855 (2.2405)  time: 2.0097  data: 0.0007  max mem: 2905
Epoch: [14]  [600/625]  eta: 0:00:50  lr: 0.002993  min_lr: 0.002993  loss: 3.9552 (3.9478)  class_acc: 0.3242 (0.3329)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8658 (2.2335)  time: 2.0167  data: 0.0008  max mem: 2905
Epoch: [14]  [624/625]  eta: 0:00:01  lr: 0.003000  min_lr: 0.003000  loss: 3.8815 (3.9471)  class_acc: 0.3359 (0.3331)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0965 (2.2444)  time: 0.6867  data: 0.0016  max mem: 2905
Epoch: [14] Total time: 0:20:22 (1.9556 s / it)
Averaged stats: lr: 0.003000  min_lr: 0.003000  loss: 3.8815 (3.9474)  class_acc: 0.3359 (0.3333)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0965 (2.2444)
Test:  [ 0/50]  eta: 0:09:58  loss: 2.2925 (2.2925)  acc1: 44.8000 (44.8000)  acc5: 78.4000 (78.4000)  time: 11.9628  data: 11.9327  max mem: 2905
Test:  [10/50]  eta: 0:01:20  loss: 2.3968 (2.4221)  acc1: 48.0000 (48.2182)  acc5: 72.8000 (72.1455)  time: 2.0113  data: 1.9916  max mem: 2905
Test:  [20/50]  eta: 0:00:45  loss: 2.5434 (2.5883)  acc1: 42.4000 (43.9619)  acc5: 67.2000 (69.7524)  time: 0.9887  data: 0.9698  max mem: 2905
Test:  [30/50]  eta: 0:00:26  loss: 2.6395 (2.5606)  acc1: 40.8000 (43.7677)  acc5: 68.8000 (70.0645)  time: 0.9109  data: 0.8925  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 2.6302 (2.5910)  acc1: 41.6000 (43.3951)  acc5: 69.6000 (69.3268)  time: 0.8205  data: 0.8026  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.6377 (2.6034)  acc1: 41.6000 (43.1360)  acc5: 67.2000 (69.1040)  time: 0.6856  data: 0.6677  max mem: 2905
Test: Total time: 0:00:52 (1.0425 s / it)
* Acc@1 44.156 Acc@5 69.362 loss 2.579
Accuracy of the model on the 50000 test images: 44.2%
Max accuracy: 44.16%
Epoch: [15]  [  0/625]  eta: 3:41:04  lr: 0.003000  min_lr: 0.003000  loss: 3.8014 (3.8014)  class_acc: 0.3711 (0.3711)  weight_decay: 0.0500 (0.0500)  time: 21.2231  data: 20.8451  max mem: 2905
Epoch: [15]  [200/625]  eta: 0:14:34  lr: 0.003065  min_lr: 0.003065  loss: 3.9123 (3.8989)  class_acc: 0.3398 (0.3444)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8742 (2.2813)  time: 1.9342  data: 0.0007  max mem: 2905
Epoch: [15]  [400/625]  eta: 0:07:34  lr: 0.003129  min_lr: 0.003129  loss: 4.0027 (3.9085)  class_acc: 0.3203 (0.3407)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6344 (2.2771)  time: 2.1218  data: 0.0007  max mem: 2905
Epoch: [15]  [600/625]  eta: 0:00:50  lr: 0.003193  min_lr: 0.003193  loss: 3.8798 (3.9093)  class_acc: 0.3477 (0.3401)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2548 (2.2944)  time: 2.0437  data: 0.0156  max mem: 2905
Epoch: [15]  [624/625]  eta: 0:00:01  lr: 0.003200  min_lr: 0.003200  loss: 3.9158 (3.9099)  class_acc: 0.3359 (0.3400)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7865 (2.2740)  time: 0.7800  data: 0.0014  max mem: 2905
Epoch: [15] Total time: 0:20:22 (1.9558 s / it)
Averaged stats: lr: 0.003200  min_lr: 0.003200  loss: 3.9158 (3.9047)  class_acc: 0.3359 (0.3413)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7865 (2.2740)
Test:  [ 0/50]  eta: 0:10:48  loss: 2.4163 (2.4163)  acc1: 52.8000 (52.8000)  acc5: 72.8000 (72.8000)  time: 12.9715  data: 12.9431  max mem: 2905
Test:  [10/50]  eta: 0:01:28  loss: 2.4163 (2.3852)  acc1: 49.6000 (49.6727)  acc5: 73.6000 (72.9455)  time: 2.2191  data: 2.1984  max mem: 2905
Test:  [20/50]  eta: 0:00:52  loss: 2.6175 (2.5718)  acc1: 44.0000 (44.9143)  acc5: 69.6000 (70.2857)  time: 1.1909  data: 1.1703  max mem: 2905
Test:  [30/50]  eta: 0:00:30  loss: 2.6738 (2.5493)  acc1: 42.4000 (44.8516)  acc5: 67.2000 (70.0645)  time: 1.1545  data: 1.1335  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 2.5792 (2.5740)  acc1: 42.4000 (44.2732)  acc5: 68.8000 (69.7756)  time: 0.7458  data: 0.7242  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.5827 (2.5912)  acc1: 42.4000 (44.1600)  acc5: 68.8000 (69.5680)  time: 0.6467  data: 0.6245  max mem: 2905
Test: Total time: 0:00:52 (1.0484 s / it)
* Acc@1 44.670 Acc@5 69.832 loss 2.567
Accuracy of the model on the 50000 test images: 44.7%
Max accuracy: 44.67%
Epoch: [16]  [  0/625]  eta: 4:06:53  lr: 0.003201  min_lr: 0.003201  loss: 3.8835 (3.8835)  class_acc: 0.3281 (0.3281)  weight_decay: 0.0500 (0.0500)  time: 23.7022  data: 23.5759  max mem: 2905
Epoch: [16]  [200/625]  eta: 0:14:28  lr: 0.003265  min_lr: 0.003265  loss: 3.8391 (3.8736)  class_acc: 0.3516 (0.3472)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2771 (2.2100)  time: 1.9181  data: 0.0462  max mem: 2905
Epoch: [16]  [400/625]  eta: 0:07:21  lr: 0.003329  min_lr: 0.003329  loss: 3.8863 (3.8710)  class_acc: 0.3477 (0.3481)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0969 (2.1265)  time: 1.9864  data: 0.1882  max mem: 2905
Epoch: [16]  [600/625]  eta: 0:00:48  lr: 0.003393  min_lr: 0.003393  loss: 3.8983 (3.8678)  class_acc: 0.3477 (0.3492)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1502 (2.1549)  time: 2.0331  data: 0.0263  max mem: 2905
Epoch: [16]  [624/625]  eta: 0:00:01  lr: 0.003400  min_lr: 0.003400  loss: 3.8396 (3.8680)  class_acc: 0.3516 (0.3491)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.6281 (2.1470)  time: 0.7000  data: 0.0013  max mem: 2905
Epoch: [16] Total time: 0:19:51 (1.9058 s / it)
Averaged stats: lr: 0.003400  min_lr: 0.003400  loss: 3.8396 (3.8673)  class_acc: 0.3516 (0.3485)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.6281 (2.1470)
Test:  [ 0/50]  eta: 0:10:17  loss: 2.2174 (2.2174)  acc1: 55.2000 (55.2000)  acc5: 76.0000 (76.0000)  time: 12.3581  data: 12.3300  max mem: 2905
Test:  [10/50]  eta: 0:01:19  loss: 2.3906 (2.3496)  acc1: 50.4000 (50.4000)  acc5: 75.2000 (73.8182)  time: 1.9922  data: 1.9721  max mem: 2905
Test:  [20/50]  eta: 0:00:45  loss: 2.4548 (2.5181)  acc1: 44.8000 (45.7524)  acc5: 71.2000 (71.4286)  time: 0.9715  data: 0.9526  max mem: 2905
Test:  [30/50]  eta: 0:00:26  loss: 2.6267 (2.5054)  acc1: 40.8000 (45.4968)  acc5: 69.6000 (71.3806)  time: 0.9471  data: 0.9276  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 2.5787 (2.5398)  acc1: 43.2000 (45.0146)  acc5: 70.4000 (70.8488)  time: 0.7296  data: 0.7099  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 2.5787 (2.5577)  acc1: 43.2000 (44.6720)  acc5: 69.6000 (70.3680)  time: 0.6492  data: 0.6307  max mem: 2905
Test: Total time: 0:00:49 (0.9956 s / it)
* Acc@1 46.040 Acc@5 71.156 loss 2.523
Accuracy of the model on the 50000 test images: 46.0%
Max accuracy: 46.04%
Epoch: [17]  [  0/625]  eta: 3:13:39  lr: 0.003401  min_lr: 0.003401  loss: 3.8910 (3.8910)  class_acc: 0.3594 (0.3594)  weight_decay: 0.0500 (0.0500)  time: 18.5909  data: 18.4678  max mem: 2905
Epoch: [17]  [200/625]  eta: 0:14:01  lr: 0.003465  min_lr: 0.003465  loss: 3.8857 (3.8389)  class_acc: 0.3477 (0.3538)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8196 (2.2447)  time: 1.9227  data: 1.6070  max mem: 2905
Epoch: [17]  [400/625]  eta: 0:07:11  lr: 0.003529  min_lr: 0.003529  loss: 3.7862 (3.8369)  class_acc: 0.3555 (0.3550)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8312 (inf)  time: 1.8386  data: 1.6031  max mem: 2905
Epoch: [17]  [600/625]  eta: 0:00:47  lr: 0.003593  min_lr: 0.003593  loss: 3.8457 (3.8430)  class_acc: 0.3516 (0.3534)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.5092 (inf)  time: 1.8567  data: 1.6188  max mem: 2905
Epoch: [17]  [624/625]  eta: 0:00:01  lr: 0.003600  min_lr: 0.003600  loss: 3.8486 (3.8432)  class_acc: 0.3477 (0.3534)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1188 (inf)  time: 0.7762  data: 0.6235  max mem: 2905
Epoch: [17] Total time: 0:19:21 (1.8588 s / it)
Averaged stats: lr: 0.003600  min_lr: 0.003600  loss: 3.8486 (3.8353)  class_acc: 0.3477 (0.3548)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1188 (inf)
Test:  [ 0/50]  eta: 0:10:14  loss: 2.5408 (2.5408)  acc1: 40.8000 (40.8000)  acc5: 71.2000 (71.2000)  time: 12.2809  data: 12.2577  max mem: 2905
Test:  [10/50]  eta: 0:01:22  loss: 2.5408 (2.5432)  acc1: 46.4000 (45.6727)  acc5: 68.8000 (69.8182)  time: 2.0601  data: 2.0388  max mem: 2905
Test:  [20/50]  eta: 0:00:48  loss: 2.7378 (2.7221)  acc1: 42.4000 (41.4095)  acc5: 67.2000 (67.0476)  time: 1.0946  data: 1.0739  max mem: 2905
Test:  [30/50]  eta: 0:00:28  loss: 2.8656 (2.7270)  acc1: 39.2000 (41.2903)  acc5: 64.0000 (66.9936)  time: 1.0673  data: 1.0479  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 2.8656 (2.7829)  acc1: 39.2000 (40.3902)  acc5: 64.0000 (65.8146)  time: 0.6944  data: 0.6748  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 2.8625 (2.7906)  acc1: 40.0000 (40.6880)  acc5: 64.0000 (65.7600)  time: 0.6024  data: 0.5829  max mem: 2905
Test: Total time: 0:00:48 (0.9757 s / it)
* Acc@1 41.648 Acc@5 66.760 loss 2.742
Accuracy of the model on the 50000 test images: 41.6%
Max accuracy: 46.04%
Epoch: [18]  [  0/625]  eta: 3:33:38  lr: 0.003601  min_lr: 0.003601  loss: 3.7695 (3.7695)  class_acc: 0.3867 (0.3867)  weight_decay: 0.0500 (0.0500)  time: 20.5104  data: 18.1215  max mem: 2905
Epoch: [18]  [200/625]  eta: 0:13:40  lr: 0.003665  min_lr: 0.003665  loss: 3.7912 (3.8065)  class_acc: 0.3633 (0.3593)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.4906 (1.9935)  time: 1.7403  data: 0.0926  max mem: 2905
Epoch: [18]  [400/625]  eta: 0:07:02  lr: 0.003729  min_lr: 0.003729  loss: 3.8296 (3.8077)  class_acc: 0.3594 (0.3597)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.6810 (2.0978)  time: 1.9347  data: 0.0006  max mem: 2905
Epoch: [18]  [600/625]  eta: 0:00:47  lr: 0.003793  min_lr: 0.003793  loss: 3.7772 (3.8113)  class_acc: 0.3438 (0.3588)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7198 (2.0771)  time: 2.0009  data: 0.0006  max mem: 2905
Epoch: [18]  [624/625]  eta: 0:00:01  lr: 0.003800  min_lr: 0.003800  loss: 3.8208 (3.8127)  class_acc: 0.3516 (0.3589)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2762 (2.1282)  time: 0.7129  data: 0.0014  max mem: 2905
Epoch: [18] Total time: 0:19:17 (1.8526 s / it)
Averaged stats: lr: 0.003800  min_lr: 0.003800  loss: 3.8208 (3.8136)  class_acc: 0.3516 (0.3592)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2762 (2.1282)
Test:  [ 0/50]  eta: 0:09:54  loss: 2.6247 (2.6247)  acc1: 45.6000 (45.6000)  acc5: 70.4000 (70.4000)  time: 11.8913  data: 11.8576  max mem: 2905
Test:  [10/50]  eta: 0:01:20  loss: 2.6708 (2.6645)  acc1: 44.0000 (43.0545)  acc5: 68.8000 (68.4364)  time: 2.0214  data: 2.0011  max mem: 2905
Test:  [20/50]  eta: 0:00:47  loss: 2.8277 (2.7971)  acc1: 38.4000 (40.0381)  acc5: 66.4000 (66.6667)  time: 1.0846  data: 1.0656  max mem: 2905
Test:  [30/50]  eta: 0:00:27  loss: 2.8516 (2.7834)  acc1: 38.4000 (40.4903)  acc5: 64.8000 (66.4000)  time: 1.0508  data: 1.0310  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 2.7912 (2.8077)  acc1: 40.8000 (39.9220)  acc5: 63.2000 (66.0293)  time: 0.7072  data: 0.6882  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 2.8734 (2.8130)  acc1: 36.8000 (39.8080)  acc5: 65.6000 (66.0320)  time: 0.6628  data: 0.6440  max mem: 2905
Test: Total time: 0:00:48 (0.9739 s / it)
* Acc@1 41.306 Acc@5 66.252 loss 2.793
Accuracy of the model on the 50000 test images: 41.3%
Max accuracy: 46.04%
Epoch: [19]  [  0/625]  eta: 3:27:37  lr: 0.003801  min_lr: 0.003801  loss: 3.7461 (3.7461)  class_acc: 0.3789 (0.3789)  weight_decay: 0.0500 (0.0500)  time: 19.9313  data: 19.0259  max mem: 2905
Epoch: [19]  [200/625]  eta: 0:13:18  lr: 0.003865  min_lr: 0.003865  loss: 3.7814 (3.7913)  class_acc: 0.3633 (0.3657)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7745 (2.0692)  time: 1.7525  data: 0.0081  max mem: 2905
Epoch: [19]  [400/625]  eta: 0:06:59  lr: 0.003929  min_lr: 0.003929  loss: 3.7692 (3.7926)  class_acc: 0.3672 (0.3659)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.5919 (2.0602)  time: 1.7423  data: 0.0006  max mem: 2905
Epoch: [19]  [600/625]  eta: 0:00:46  lr: 0.003993  min_lr: 0.003993  loss: 3.7743 (3.7874)  class_acc: 0.3672 (0.3665)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8381 (2.0593)  time: 1.7971  data: 0.0006  max mem: 2905
Epoch: [19]  [624/625]  eta: 0:00:01  lr: 0.004000  min_lr: 0.004000  loss: 3.7813 (3.7878)  class_acc: 0.3633 (0.3663)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7304 (2.0414)  time: 0.9912  data: 0.0013  max mem: 2905
Epoch: [19] Total time: 0:19:13 (1.8450 s / it)
Averaged stats: lr: 0.004000  min_lr: 0.004000  loss: 3.7813 (3.7908)  class_acc: 0.3633 (0.3636)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7304 (2.0414)
Test:  [ 0/50]  eta: 0:10:33  loss: 2.4847 (2.4847)  acc1: 39.2000 (39.2000)  acc5: 69.6000 (69.6000)  time: 12.6661  data: 12.6366  max mem: 2905
Test:  [10/50]  eta: 0:01:30  loss: 2.3167 (2.3055)  acc1: 51.2000 (48.5818)  acc5: 73.6000 (74.7636)  time: 2.2625  data: 2.2422  max mem: 2905
Test:  [20/50]  eta: 0:00:54  loss: 2.5198 (2.4631)  acc1: 45.6000 (45.0286)  acc5: 72.0000 (72.3048)  time: 1.2663  data: 1.2454  max mem: 2905
Test:  [30/50]  eta: 0:00:29  loss: 2.6153 (2.4630)  acc1: 43.2000 (45.3161)  acc5: 69.6000 (71.8194)  time: 1.0732  data: 1.0521  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 2.5907 (2.4913)  acc1: 44.8000 (45.3659)  acc5: 68.0000 (71.2781)  time: 0.5919  data: 0.5697  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.6554 (2.5000)  acc1: 44.8000 (45.2480)  acc5: 68.0000 (70.8800)  time: 0.5238  data: 0.5017  max mem: 2905
Test: Total time: 0:00:50 (1.0104 s / it)
* Acc@1 46.238 Acc@5 71.376 loss 2.472
Accuracy of the model on the 50000 test images: 46.2%
Max accuracy: 46.24%
Epoch: [20]  [  0/625]  eta: 3:41:23  lr: 0.004000  min_lr: 0.004000  loss: 3.5793 (3.5793)  class_acc: 0.3906 (0.3906)  weight_decay: 0.0500 (0.0500)  time: 21.2543  data: 16.9177  max mem: 2905
Epoch: [20]  [200/625]  eta: 0:14:01  lr: 0.004000  min_lr: 0.004000  loss: 3.7860 (3.7508)  class_acc: 0.3516 (0.3733)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7203 (1.9910)  time: 1.8292  data: 0.0011  max mem: 2905
Epoch: [20]  [400/625]  eta: 0:07:15  lr: 0.004000  min_lr: 0.004000  loss: 3.7795 (3.7539)  class_acc: 0.3711 (0.3730)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7622 (2.0548)  time: 1.9608  data: 0.0008  max mem: 2905
Epoch: [20]  [600/625]  eta: 0:00:48  lr: 0.004000  min_lr: 0.004000  loss: 3.7585 (3.7597)  class_acc: 0.3672 (0.3716)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.5971 (2.0682)  time: 1.8815  data: 0.0011  max mem: 2905
Epoch: [20]  [624/625]  eta: 0:00:01  lr: 0.004000  min_lr: 0.004000  loss: 3.8036 (3.7596)  class_acc: 0.3633 (0.3715)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7657 (2.0750)  time: 0.7967  data: 0.0013  max mem: 2905
Epoch: [20] Total time: 0:19:32 (1.8759 s / it)
Averaged stats: lr: 0.004000  min_lr: 0.004000  loss: 3.8036 (3.7625)  class_acc: 0.3633 (0.3694)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7657 (2.0750)
Test:  [ 0/50]  eta: 0:10:29  loss: 3.0842 (3.0842)  acc1: 32.8000 (32.8000)  acc5: 60.8000 (60.8000)  time: 12.5883  data: 12.5592  max mem: 2905
Test:  [10/50]  eta: 0:01:21  loss: 2.8227 (2.7515)  acc1: 44.0000 (41.7455)  acc5: 65.6000 (67.2000)  time: 2.0425  data: 2.0229  max mem: 2905
Test:  [20/50]  eta: 0:00:46  loss: 2.8696 (2.9127)  acc1: 39.2000 (38.8571)  acc5: 64.8000 (65.2571)  time: 0.9919  data: 0.9734  max mem: 2905
Test:  [30/50]  eta: 0:00:27  loss: 3.0171 (2.8860)  acc1: 36.8000 (39.7419)  acc5: 62.4000 (64.8258)  time: 0.9784  data: 0.9600  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 2.8158 (2.9152)  acc1: 36.8000 (38.9073)  acc5: 61.6000 (64.4293)  time: 0.7555  data: 0.7357  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.8438 (2.9174)  acc1: 36.8000 (38.8960)  acc5: 64.8000 (64.5600)  time: 0.5881  data: 0.5677  max mem: 2905
Test: Total time: 0:00:50 (1.0061 s / it)
* Acc@1 39.712 Acc@5 64.622 loss 2.871
Accuracy of the model on the 50000 test images: 39.7%
Max accuracy: 46.24%
Epoch: [21]  [  0/625]  eta: 4:05:07  lr: 0.004000  min_lr: 0.004000  loss: 3.7904 (3.7904)  class_acc: 0.3672 (0.3672)  weight_decay: 0.0500 (0.0500)  time: 23.5315  data: 17.3222  max mem: 2905
Epoch: [21]  [200/625]  eta: 0:14:13  lr: 0.004000  min_lr: 0.004000  loss: 3.6553 (3.7189)  class_acc: 0.3828 (0.3741)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8959 (1.9880)  time: 2.0153  data: 0.0016  max mem: 2905
Epoch: [21]  [400/625]  eta: 0:07:19  lr: 0.004000  min_lr: 0.004000  loss: 3.6694 (3.7235)  class_acc: 0.3750 (0.3756)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2716 (2.1337)  time: 1.9323  data: 0.0006  max mem: 2905
Epoch: [21]  [600/625]  eta: 0:00:48  lr: 0.004000  min_lr: 0.004000  loss: 3.7389 (3.7282)  class_acc: 0.3750 (0.3741)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7289 (2.0225)  time: 1.9361  data: 0.0006  max mem: 2905
Epoch: [21]  [624/625]  eta: 0:00:01  lr: 0.003999  min_lr: 0.003999  loss: 3.6977 (3.7290)  class_acc: 0.3750 (0.3740)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8270 (2.0189)  time: 0.8300  data: 0.0016  max mem: 2905
Epoch: [21] Total time: 0:19:52 (1.9078 s / it)
Averaged stats: lr: 0.003999  min_lr: 0.003999  loss: 3.6977 (3.7321)  class_acc: 0.3750 (0.3755)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8270 (2.0189)
Test:  [ 0/50]  eta: 0:10:33  loss: 2.4033 (2.4033)  acc1: 46.4000 (46.4000)  acc5: 70.4000 (70.4000)  time: 12.6693  data: 12.6438  max mem: 2905
Test:  [10/50]  eta: 0:01:30  loss: 2.3841 (2.4315)  acc1: 47.2000 (47.6364)  acc5: 71.2000 (70.9818)  time: 2.2617  data: 2.2409  max mem: 2905
Test:  [20/50]  eta: 0:00:53  loss: 2.5151 (2.5363)  acc1: 46.4000 (45.3333)  acc5: 71.2000 (70.4381)  time: 1.2458  data: 1.2264  max mem: 2905
Test:  [30/50]  eta: 0:00:29  loss: 2.6047 (2.5368)  acc1: 44.0000 (45.2129)  acc5: 70.4000 (70.3226)  time: 1.0357  data: 1.0170  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 2.6984 (2.5929)  acc1: 43.2000 (44.3317)  acc5: 68.0000 (69.2293)  time: 0.5806  data: 0.5621  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 2.5740 (2.5983)  acc1: 44.0000 (44.4480)  acc5: 68.8000 (69.2640)  time: 0.5218  data: 0.5029  max mem: 2905
Test: Total time: 0:00:50 (1.0020 s / it)
* Acc@1 44.660 Acc@5 69.776 loss 2.574
Accuracy of the model on the 50000 test images: 44.7%
Max accuracy: 46.24%
Epoch: [22]  [  0/625]  eta: 3:27:10  lr: 0.003999  min_lr: 0.003999  loss: 3.8757 (3.8757)  class_acc: 0.3516 (0.3516)  weight_decay: 0.0500 (0.0500)  time: 19.8884  data: 18.6958  max mem: 2905
Epoch: [22]  [200/625]  eta: 0:13:47  lr: 0.003999  min_lr: 0.003999  loss: 3.7881 (3.7122)  class_acc: 0.3672 (0.3787)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7065 (1.9623)  time: 1.8335  data: 0.0006  max mem: 2905
Epoch: [22]  [400/625]  eta: 0:07:11  lr: 0.003999  min_lr: 0.003999  loss: 3.6754 (3.7072)  class_acc: 0.3750 (0.3799)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7173 (1.9558)  time: 1.8211  data: 0.0008  max mem: 2905
Epoch: [22]  [600/625]  eta: 0:00:47  lr: 0.003999  min_lr: 0.003999  loss: 3.6931 (3.7025)  class_acc: 0.3828 (0.3805)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7012 (1.9935)  time: 1.7528  data: 0.0110  max mem: 2905
Epoch: [22]  [624/625]  eta: 0:00:01  lr: 0.003999  min_lr: 0.003999  loss: 3.6958 (3.7020)  class_acc: 0.3906 (0.3807)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8089 (1.9925)  time: 0.8893  data: 0.0082  max mem: 2905
Epoch: [22] Total time: 0:19:43 (1.8940 s / it)
Averaged stats: lr: 0.003999  min_lr: 0.003999  loss: 3.6958 (3.7013)  class_acc: 0.3906 (0.3812)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8089 (1.9925)
Test:  [ 0/50]  eta: 0:09:19  loss: 2.1642 (2.1642)  acc1: 53.6000 (53.6000)  acc5: 73.6000 (73.6000)  time: 11.1931  data: 11.1673  max mem: 2905
Test:  [10/50]  eta: 0:01:08  loss: 2.2309 (2.1733)  acc1: 53.6000 (53.8182)  acc5: 75.2000 (76.0000)  time: 1.7047  data: 1.6841  max mem: 2905
Test:  [20/50]  eta: 0:00:41  loss: 2.2977 (2.3197)  acc1: 49.6000 (49.9048)  acc5: 72.8000 (74.2857)  time: 0.8950  data: 0.8757  max mem: 2905
Test:  [30/50]  eta: 0:00:26  loss: 2.4963 (2.3411)  acc1: 45.6000 (49.4452)  acc5: 72.0000 (73.4710)  time: 1.0874  data: 1.0689  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 2.4343 (2.3794)  acc1: 47.2000 (48.4293)  acc5: 71.2000 (72.8390)  time: 0.9650  data: 0.9471  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.3533 (2.3700)  acc1: 48.0000 (48.4800)  acc5: 72.0000 (73.0880)  time: 0.5751  data: 0.5571  max mem: 2905
Test: Total time: 0:00:52 (1.0429 s / it)
* Acc@1 48.866 Acc@5 73.468 loss 2.344
Accuracy of the model on the 50000 test images: 48.9%
Max accuracy: 48.87%
Epoch: [23]  [  0/625]  eta: 3:39:53  lr: 0.003999  min_lr: 0.003999  loss: 3.5973 (3.5973)  class_acc: 0.4336 (0.4336)  weight_decay: 0.0500 (0.0500)  time: 21.1095  data: 20.3723  max mem: 2905
Epoch: [23]  [200/625]  eta: 0:14:14  lr: 0.003999  min_lr: 0.003999  loss: 3.6491 (3.6711)  class_acc: 0.3789 (0.3890)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9880 (2.2326)  time: 1.9598  data: 0.6415  max mem: 2905
Epoch: [23]  [400/625]  eta: 0:07:16  lr: 0.003998  min_lr: 0.003998  loss: 3.7368 (3.6840)  class_acc: 0.3750 (0.3860)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.6813 (2.1713)  time: 1.8327  data: 0.0009  max mem: 2905
Epoch: [23]  [600/625]  eta: 0:00:48  lr: 0.003998  min_lr: 0.003998  loss: 3.6761 (3.6815)  class_acc: 0.3789 (0.3858)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.5312 (inf)  time: 1.8845  data: 0.0417  max mem: 2905
Epoch: [23]  [624/625]  eta: 0:00:01  lr: 0.003998  min_lr: 0.003998  loss: 3.6482 (3.6813)  class_acc: 0.3789 (0.3858)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7138 (inf)  time: 0.8022  data: 0.0152  max mem: 2905
Epoch: [23] Total time: 0:19:34 (1.8799 s / it)
Averaged stats: lr: 0.003998  min_lr: 0.003998  loss: 3.6482 (3.6764)  class_acc: 0.3789 (0.3861)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7138 (inf)
Test:  [ 0/50]  eta: 0:10:17  loss: 2.3520 (2.3520)  acc1: 49.6000 (49.6000)  acc5: 72.8000 (72.8000)  time: 12.3509  data: 12.3117  max mem: 2905
Test:  [10/50]  eta: 0:01:09  loss: 2.3520 (2.3535)  acc1: 48.8000 (48.9455)  acc5: 72.0000 (73.0909)  time: 1.7484  data: 1.7276  max mem: 2905
Test:  [20/50]  eta: 0:00:38  loss: 2.5317 (2.5242)  acc1: 44.0000 (45.4476)  acc5: 71.2000 (70.1714)  time: 0.7383  data: 0.7184  max mem: 2905
Test:  [30/50]  eta: 0:00:24  loss: 2.6113 (2.5221)  acc1: 43.2000 (45.2903)  acc5: 68.0000 (70.1419)  time: 0.9622  data: 0.9422  max mem: 2905
Test:  [40/50]  eta: 0:00:10  loss: 2.6022 (2.5431)  acc1: 43.2000 (44.9561)  acc5: 68.0000 (70.0488)  time: 0.8729  data: 0.8520  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 2.6396 (2.5664)  acc1: 43.2000 (44.6560)  acc5: 68.8000 (69.9200)  time: 0.4563  data: 0.4363  max mem: 2905
Test: Total time: 0:00:47 (0.9547 s / it)
* Acc@1 45.160 Acc@5 70.234 loss 2.548
Accuracy of the model on the 50000 test images: 45.2%
Max accuracy: 48.87%
Epoch: [24]  [  0/625]  eta: 4:03:15  lr: 0.003998  min_lr: 0.003998  loss: 3.6114 (3.6114)  class_acc: 0.4141 (0.4141)  weight_decay: 0.0500 (0.0500)  time: 23.3534  data: 19.9394  max mem: 2905
Epoch: [24]  [200/625]  eta: 0:14:07  lr: 0.003998  min_lr: 0.003998  loss: 3.6764 (3.6450)  class_acc: 0.3828 (0.3905)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.4303 (2.1383)  time: 1.8211  data: 0.0007  max mem: 2905
Epoch: [24]  [400/625]  eta: 0:07:12  lr: 0.003997  min_lr: 0.003997  loss: 3.6918 (3.6442)  class_acc: 0.3789 (0.3912)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.6180 (2.1398)  time: 1.8983  data: 0.0186  max mem: 2905
Epoch: [24]  [600/625]  eta: 0:00:48  lr: 0.003997  min_lr: 0.003997  loss: 3.5939 (3.6418)  class_acc: 0.3828 (0.3918)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8641 (2.1125)  time: 2.1217  data: 0.0462  max mem: 2905
Epoch: [24]  [624/625]  eta: 0:00:01  lr: 0.003997  min_lr: 0.003997  loss: 3.6427 (3.6419)  class_acc: 0.3906 (0.3918)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.6812 (2.1027)  time: 0.6184  data: 0.0013  max mem: 2905
Epoch: [24] Total time: 0:20:08 (1.9335 s / it)
Averaged stats: lr: 0.003997  min_lr: 0.003997  loss: 3.6427 (3.6528)  class_acc: 0.3906 (0.3904)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.6812 (2.1027)
Test:  [ 0/50]  eta: 0:11:01  loss: 2.1656 (2.1656)  acc1: 52.8000 (52.8000)  acc5: 79.2000 (79.2000)  time: 13.2280  data: 13.1921  max mem: 2905
Test:  [10/50]  eta: 0:01:27  loss: 2.1610 (2.1598)  acc1: 54.4000 (54.2545)  acc5: 78.4000 (76.4364)  time: 2.1764  data: 2.1563  max mem: 2905
Test:  [20/50]  eta: 0:00:51  loss: 2.3438 (2.3316)  acc1: 50.4000 (49.6381)  acc5: 73.6000 (74.5143)  time: 1.1365  data: 1.1169  max mem: 2905
Test:  [30/50]  eta: 0:00:30  loss: 2.4677 (2.3443)  acc1: 44.8000 (48.9290)  acc5: 72.8000 (74.0645)  time: 1.1834  data: 1.1620  max mem: 2905
Test:  [40/50]  eta: 0:00:13  loss: 2.3905 (2.3670)  acc1: 46.4000 (48.8976)  acc5: 72.0000 (73.4634)  time: 0.9036  data: 0.8819  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.3807 (2.3688)  acc1: 48.8000 (49.2160)  acc5: 72.0000 (73.4560)  time: 0.8287  data: 0.8077  max mem: 2905
Test: Total time: 0:00:54 (1.0976 s / it)
* Acc@1 48.894 Acc@5 73.396 loss 2.367
Accuracy of the model on the 50000 test images: 48.9%
Max accuracy: 48.89%
Epoch: [25]  [  0/625]  eta: 4:01:03  lr: 0.003997  min_lr: 0.003997  loss: 3.5686 (3.5686)  class_acc: 0.3945 (0.3945)  weight_decay: 0.0500 (0.0500)  time: 23.1420  data: 22.6885  max mem: 2905
Epoch: [25]  [200/625]  eta: 0:14:33  lr: 0.003996  min_lr: 0.003996  loss: 3.6051 (3.6239)  class_acc: 0.4023 (0.3945)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9610 (2.0712)  time: 1.9828  data: 0.0008  max mem: 2905
Epoch: [25]  [400/625]  eta: 0:07:28  lr: 0.003996  min_lr: 0.003996  loss: 3.6724 (3.6338)  class_acc: 0.3750 (0.3932)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1317 (2.0402)  time: 1.9872  data: 0.0009  max mem: 2905
Epoch: [25]  [600/625]  eta: 0:00:49  lr: 0.003996  min_lr: 0.003996  loss: 3.6125 (3.6347)  class_acc: 0.3711 (0.3930)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1590 (2.0136)  time: 1.9020  data: 0.0012  max mem: 2905
Epoch: [25]  [624/625]  eta: 0:00:01  lr: 0.003995  min_lr: 0.003995  loss: 3.6628 (3.6352)  class_acc: 0.3867 (0.3928)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7407 (2.0130)  time: 0.7558  data: 0.0026  max mem: 2905
Epoch: [25] Total time: 0:20:02 (1.9239 s / it)
Averaged stats: lr: 0.003995  min_lr: 0.003995  loss: 3.6628 (3.6307)  class_acc: 0.3867 (0.3951)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7407 (2.0130)
Test:  [ 0/50]  eta: 0:09:56  loss: 2.2252 (2.2252)  acc1: 52.0000 (52.0000)  acc5: 78.4000 (78.4000)  time: 11.9393  data: 11.9016  max mem: 2905
Test:  [10/50]  eta: 0:01:23  loss: 2.3495 (2.3328)  acc1: 49.6000 (50.1091)  acc5: 73.6000 (73.4545)  time: 2.0886  data: 2.0668  max mem: 2905
Test:  [20/50]  eta: 0:00:50  loss: 2.4747 (2.4821)  acc1: 47.2000 (46.3238)  acc5: 71.2000 (72.1143)  time: 1.1564  data: 1.1366  max mem: 2905
Test:  [30/50]  eta: 0:00:28  loss: 2.5827 (2.4736)  acc1: 43.2000 (46.0645)  acc5: 69.6000 (72.1032)  time: 1.0338  data: 1.0141  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 2.5845 (2.4907)  acc1: 44.8000 (46.1463)  acc5: 68.0000 (71.5902)  time: 0.5820  data: 0.5610  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 2.4665 (2.4988)  acc1: 45.6000 (46.0480)  acc5: 71.2000 (71.3280)  time: 0.4792  data: 0.4575  max mem: 2905
Test: Total time: 0:00:47 (0.9536 s / it)
* Acc@1 46.482 Acc@5 71.380 loss 2.475
Accuracy of the model on the 50000 test images: 46.5%
Max accuracy: 48.89%
Epoch: [26]  [  0/625]  eta: 3:26:09  lr: 0.003995  min_lr: 0.003995  loss: 3.7965 (3.7965)  class_acc: 0.3867 (0.3867)  weight_decay: 0.0500 (0.0500)  time: 19.7910  data: 18.9403  max mem: 2905
Epoch: [26]  [200/625]  eta: 0:13:49  lr: 0.003995  min_lr: 0.003995  loss: 3.6161 (3.5845)  class_acc: 0.4062 (0.4035)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.5103 (2.1486)  time: 2.0134  data: 0.7893  max mem: 2905
Epoch: [26]  [400/625]  eta: 0:07:11  lr: 0.003994  min_lr: 0.003994  loss: 3.5623 (3.5993)  class_acc: 0.4023 (0.4004)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.4205 (2.0950)  time: 1.9136  data: 0.5641  max mem: 2905
Epoch: [26]  [600/625]  eta: 0:00:48  lr: 0.003994  min_lr: 0.003994  loss: 3.5630 (3.6026)  class_acc: 0.3984 (0.3999)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7351 (2.1099)  time: 1.8916  data: 0.0527  max mem: 2905
Epoch: [26]  [624/625]  eta: 0:00:01  lr: 0.003994  min_lr: 0.003994  loss: 3.5775 (3.6028)  class_acc: 0.3984 (0.3997)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.6100 (2.0984)  time: 0.6861  data: 0.0291  max mem: 2905
Epoch: [26] Total time: 0:19:50 (1.9047 s / it)
Averaged stats: lr: 0.003994  min_lr: 0.003994  loss: 3.5775 (3.6107)  class_acc: 0.3984 (0.3985)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.6100 (2.0984)
Test:  [ 0/50]  eta: 0:10:45  loss: 2.3278 (2.3278)  acc1: 48.8000 (48.8000)  acc5: 74.4000 (74.4000)  time: 12.9197  data: 12.8940  max mem: 2905
Test:  [10/50]  eta: 0:01:25  loss: 2.3480 (2.3127)  acc1: 48.8000 (49.0182)  acc5: 74.4000 (74.4000)  time: 2.1292  data: 2.1102  max mem: 2905
Test:  [20/50]  eta: 0:00:49  loss: 2.4423 (2.4436)  acc1: 46.4000 (46.5524)  acc5: 71.2000 (72.4952)  time: 1.0935  data: 1.0744  max mem: 2905
Test:  [30/50]  eta: 0:00:29  loss: 2.5074 (2.4321)  acc1: 45.6000 (46.9936)  acc5: 70.4000 (72.2839)  time: 1.1002  data: 1.0800  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 2.4564 (2.4516)  acc1: 45.6000 (46.9268)  acc5: 68.8000 (71.8439)  time: 0.7694  data: 0.7475  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.5118 (2.4673)  acc1: 44.8000 (46.4320)  acc5: 68.0000 (71.2640)  time: 0.7267  data: 0.7061  max mem: 2905
Test: Total time: 0:00:50 (1.0174 s / it)
* Acc@1 47.270 Acc@5 72.132 loss 2.443
Accuracy of the model on the 50000 test images: 47.3%
Max accuracy: 48.89%
Epoch: [27]  [  0/625]  eta: 3:49:47  lr: 0.003994  min_lr: 0.003994  loss: 3.5339 (3.5339)  class_acc: 0.4141 (0.4141)  weight_decay: 0.0500 (0.0500)  time: 22.0595  data: 15.4624  max mem: 2905
Epoch: [27]  [200/625]  eta: 0:14:15  lr: 0.003993  min_lr: 0.003993  loss: 3.5980 (3.5782)  class_acc: 0.3984 (0.4067)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2039 (inf)  time: 1.8993  data: 0.0070  max mem: 2905
Epoch: [27]  [400/625]  eta: 0:07:14  lr: 0.003993  min_lr: 0.003993  loss: 3.6139 (3.5753)  class_acc: 0.3867 (0.4067)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8785 (inf)  time: 1.8946  data: 0.0008  max mem: 2905
Epoch: [27]  [600/625]  eta: 0:00:47  lr: 0.003992  min_lr: 0.003992  loss: 3.6045 (3.5791)  class_acc: 0.3984 (0.4064)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.4802 (inf)  time: 2.0196  data: 0.0316  max mem: 2905
Epoch: [27]  [624/625]  eta: 0:00:01  lr: 0.003992  min_lr: 0.003992  loss: 3.5842 (3.5794)  class_acc: 0.3867 (0.4060)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.4526 (inf)  time: 0.3925  data: 0.0258  max mem: 2905
Epoch: [27] Total time: 0:19:41 (1.8902 s / it)
Averaged stats: lr: 0.003992  min_lr: 0.003992  loss: 3.5842 (3.5927)  class_acc: 0.3867 (0.4023)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.4526 (inf)
Test:  [ 0/50]  eta: 0:10:33  loss: 2.0578 (2.0578)  acc1: 54.4000 (54.4000)  acc5: 79.2000 (79.2000)  time: 12.6712  data: 12.6441  max mem: 2905
Test:  [10/50]  eta: 0:01:26  loss: 2.1337 (2.1419)  acc1: 54.4000 (53.8182)  acc5: 76.0000 (76.3636)  time: 2.1702  data: 2.1514  max mem: 2905
Test:  [20/50]  eta: 0:00:52  loss: 2.2473 (2.2373)  acc1: 51.2000 (51.1238)  acc5: 76.0000 (75.6191)  time: 1.1922  data: 1.1741  max mem: 2905
Test:  [30/50]  eta: 0:00:30  loss: 2.2980 (2.2494)  acc1: 48.0000 (50.6065)  acc5: 74.4000 (75.2000)  time: 1.1977  data: 1.1790  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 2.2357 (2.2689)  acc1: 47.2000 (50.3415)  acc5: 72.8000 (74.7707)  time: 0.7709  data: 0.7522  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.2537 (2.2768)  acc1: 51.2000 (50.1760)  acc5: 73.6000 (74.5120)  time: 0.7491  data: 0.7296  max mem: 2905
Test: Total time: 0:00:52 (1.0513 s / it)
* Acc@1 50.538 Acc@5 75.020 loss 2.257
Accuracy of the model on the 50000 test images: 50.5%
Max accuracy: 50.54%
Epoch: [28]  [  0/625]  eta: 3:31:58  lr: 0.003992  min_lr: 0.003992  loss: 3.6734 (3.6734)  class_acc: 0.3750 (0.3750)  weight_decay: 0.0500 (0.0500)  time: 20.3503  data: 20.0512  max mem: 2905
Epoch: [28]  [200/625]  eta: 0:14:28  lr: 0.003991  min_lr: 0.003991  loss: 3.6027 (3.5692)  class_acc: 0.3906 (0.4049)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.6694 (2.1667)  time: 1.9741  data: 0.4945  max mem: 2905
Epoch: [28]  [400/625]  eta: 0:07:31  lr: 0.003991  min_lr: 0.003991  loss: 3.5839 (3.5724)  class_acc: 0.4141 (0.4070)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7739 (2.0792)  time: 2.1572  data: 0.0009  max mem: 2905
Epoch: [28]  [600/625]  eta: 0:00:49  lr: 0.003990  min_lr: 0.003990  loss: 3.5119 (3.5714)  class_acc: 0.4219 (0.4068)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7313 (2.2045)  time: 1.8744  data: 0.0008  max mem: 2905
Epoch: [28]  [624/625]  eta: 0:00:01  lr: 0.003990  min_lr: 0.003990  loss: 3.5770 (3.5711)  class_acc: 0.4102 (0.4070)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.5188 (2.1853)  time: 0.8201  data: 0.0022  max mem: 2905
Epoch: [28] Total time: 0:20:14 (1.9439 s / it)
Averaged stats: lr: 0.003990  min_lr: 0.003990  loss: 3.5770 (3.5763)  class_acc: 0.4102 (0.4058)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.5188 (2.1853)
Test:  [ 0/50]  eta: 0:10:08  loss: 2.6075 (2.6075)  acc1: 41.6000 (41.6000)  acc5: 72.0000 (72.0000)  time: 12.1648  data: 12.1361  max mem: 2905
Test:  [10/50]  eta: 0:01:18  loss: 2.3099 (2.2855)  acc1: 52.0000 (51.3455)  acc5: 73.6000 (74.2545)  time: 1.9731  data: 1.9537  max mem: 2905
Test:  [20/50]  eta: 0:00:45  loss: 2.4136 (2.4097)  acc1: 51.2000 (48.9905)  acc5: 72.0000 (72.5714)  time: 0.9957  data: 0.9773  max mem: 2905
Test:  [30/50]  eta: 0:00:26  loss: 2.4623 (2.4255)  acc1: 45.6000 (47.8968)  acc5: 71.2000 (72.2323)  time: 0.9975  data: 0.9782  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 2.4623 (2.4689)  acc1: 44.0000 (46.8683)  acc5: 69.6000 (71.4537)  time: 0.7210  data: 0.7008  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 2.5017 (2.4676)  acc1: 44.0000 (46.8640)  acc5: 72.0000 (71.5840)  time: 0.6412  data: 0.6217  max mem: 2905
Test: Total time: 0:00:48 (0.9798 s / it)
* Acc@1 47.210 Acc@5 72.040 loss 2.436
Accuracy of the model on the 50000 test images: 47.2%
Max accuracy: 50.54%
Epoch: [29]  [  0/625]  eta: 3:32:55  lr: 0.003990  min_lr: 0.003990  loss: 3.7204 (3.7204)  class_acc: 0.3633 (0.3633)  weight_decay: 0.0500 (0.0500)  time: 20.4412  data: 17.6682  max mem: 2905
Epoch: [29]  [200/625]  eta: 0:13:53  lr: 0.003989  min_lr: 0.003989  loss: 3.6205 (3.5465)  class_acc: 0.4062 (0.4141)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8717 (2.3388)  time: 1.8569  data: 0.0008  max mem: 2905
Epoch: [29]  [400/625]  eta: 0:07:16  lr: 0.003988  min_lr: 0.003988  loss: 3.6023 (3.5596)  class_acc: 0.4102 (0.4110)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.2989 (2.1954)  time: 1.9144  data: 0.0195  max mem: 2905
Epoch: [29]  [600/625]  eta: 0:00:49  lr: 0.003988  min_lr: 0.003988  loss: 3.5655 (3.5613)  class_acc: 0.3945 (0.4096)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.4800 (2.1547)  time: 2.0340  data: 0.0050  max mem: 2905
Epoch: [29]  [624/625]  eta: 0:00:01  lr: 0.003987  min_lr: 0.003987  loss: 3.5300 (3.5614)  class_acc: 0.4102 (0.4095)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.5709 (2.1348)  time: 0.4527  data: 0.0014  max mem: 2905
Epoch: [29] Total time: 0:20:05 (1.9282 s / it)
Averaged stats: lr: 0.003987  min_lr: 0.003987  loss: 3.5300 (3.5587)  class_acc: 0.4102 (0.4088)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.5709 (2.1348)
Test:  [ 0/50]  eta: 0:10:31  loss: 2.3556 (2.3556)  acc1: 48.8000 (48.8000)  acc5: 73.6000 (73.6000)  time: 12.6234  data: 12.5951  max mem: 2905
Test:  [10/50]  eta: 0:01:20  loss: 2.3297 (2.3093)  acc1: 51.2000 (51.0545)  acc5: 73.6000 (74.6909)  time: 2.0217  data: 1.9997  max mem: 2905
Test:  [20/50]  eta: 0:00:46  loss: 2.5162 (2.4935)  acc1: 45.6000 (46.8191)  acc5: 71.2000 (72.0000)  time: 1.0032  data: 0.9816  max mem: 2905
Test:  [30/50]  eta: 0:00:27  loss: 2.5584 (2.4914)  acc1: 43.2000 (46.4516)  acc5: 68.8000 (71.2774)  time: 1.0244  data: 1.0035  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 2.5514 (2.5317)  acc1: 43.2000 (45.9902)  acc5: 67.2000 (70.6927)  time: 0.9106  data: 0.8909  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.5967 (2.5288)  acc1: 44.8000 (46.1600)  acc5: 70.4000 (71.0560)  time: 0.8203  data: 0.8014  max mem: 2905
Test: Total time: 0:00:54 (1.0977 s / it)
* Acc@1 46.138 Acc@5 71.058 loss 2.510
Accuracy of the model on the 50000 test images: 46.1%
Max accuracy: 50.54%
Epoch: [30]  [  0/625]  eta: 3:25:37  lr: 0.003987  min_lr: 0.003987  loss: 3.2437 (3.2437)  class_acc: 0.4492 (0.4492)  weight_decay: 0.0500 (0.0500)  time: 19.7393  data: 17.9200  max mem: 2905
Epoch: [30]  [200/625]  eta: 0:14:18  lr: 0.003987  min_lr: 0.003987  loss: 3.5421 (3.5261)  class_acc: 0.4102 (0.4111)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3030 (2.4305)  time: 1.7800  data: 0.0008  max mem: 2905
Epoch: [30]  [400/625]  eta: 0:07:21  lr: 0.003986  min_lr: 0.003986  loss: 3.5666 (3.5352)  class_acc: 0.4141 (0.4122)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.6741 (2.2632)  time: 1.9781  data: 0.0007  max mem: 2905
Epoch: [30]  [600/625]  eta: 0:00:49  lr: 0.003985  min_lr: 0.003985  loss: 3.5600 (3.5398)  class_acc: 0.3984 (0.4118)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0842 (2.2458)  time: 2.2599  data: 0.0626  max mem: 2905
Epoch: [30]  [624/625]  eta: 0:00:01  lr: 0.003985  min_lr: 0.003985  loss: 3.5901 (3.5410)  class_acc: 0.3906 (0.4114)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8053 (2.2553)  time: 1.1417  data: 0.0063  max mem: 2905
Epoch: [30] Total time: 0:20:03 (1.9257 s / it)
Averaged stats: lr: 0.003985  min_lr: 0.003985  loss: 3.5901 (3.5421)  class_acc: 0.3906 (0.4121)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8053 (2.2553)
Test:  [ 0/50]  eta: 0:10:32  loss: 2.7928 (2.7928)  acc1: 40.8000 (40.8000)  acc5: 67.2000 (67.2000)  time: 12.6565  data: 12.6298  max mem: 2905
Test:  [10/50]  eta: 0:01:27  loss: 2.6436 (2.5998)  acc1: 47.2000 (45.9636)  acc5: 68.8000 (70.1091)  time: 2.1983  data: 2.1711  max mem: 2905
Test:  [20/50]  eta: 0:00:49  loss: 2.7254 (2.7817)  acc1: 40.0000 (41.5619)  acc5: 67.2000 (67.2000)  time: 1.0826  data: 1.0593  max mem: 2905
Test:  [30/50]  eta: 0:00:27  loss: 2.9964 (2.8042)  acc1: 36.8000 (40.9806)  acc5: 63.2000 (66.3484)  time: 0.8918  data: 0.8714  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 2.9047 (2.8522)  acc1: 36.8000 (40.0000)  acc5: 61.6000 (65.4244)  time: 0.5747  data: 0.5550  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 2.8493 (2.8554)  acc1: 36.0000 (39.9040)  acc5: 64.8000 (65.4400)  time: 0.4939  data: 0.4732  max mem: 2905
Test: Total time: 0:00:46 (0.9399 s / it)
* Acc@1 40.982 Acc@5 65.704 loss 2.836
Accuracy of the model on the 50000 test images: 41.0%
Max accuracy: 50.54%
Epoch: [31]  [  0/625]  eta: 3:53:25  lr: 0.003985  min_lr: 0.003985  loss: 3.6486 (3.6486)  class_acc: 0.4297 (0.4297)  weight_decay: 0.0500 (0.0500)  time: 22.4095  data: 20.3318  max mem: 2905
Epoch: [31]  [200/625]  eta: 0:14:04  lr: 0.003984  min_lr: 0.003984  loss: 3.6013 (3.5261)  class_acc: 0.4023 (0.4139)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3462 (1.8529)  time: 1.7383  data: 0.1542  max mem: 2905
Epoch: [31]  [400/625]  eta: 0:07:14  lr: 0.003983  min_lr: 0.003983  loss: 3.5149 (3.5341)  class_acc: 0.4219 (0.4127)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7822 (2.0280)  time: 1.7648  data: 0.8956  max mem: 2905
Epoch: [31]  [600/625]  eta: 0:00:48  lr: 0.003982  min_lr: 0.003982  loss: 3.5508 (3.5355)  class_acc: 0.4102 (0.4137)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.6879 (2.0432)  time: 1.8432  data: 0.0007  max mem: 2905
Epoch: [31]  [624/625]  eta: 0:00:01  lr: 0.003982  min_lr: 0.003982  loss: 3.5453 (3.5359)  class_acc: 0.4141 (0.4137)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.4899 (2.0233)  time: 0.8732  data: 0.0015  max mem: 2905
Epoch: [31] Total time: 0:19:48 (1.9012 s / it)
Averaged stats: lr: 0.003982  min_lr: 0.003982  loss: 3.5453 (3.5250)  class_acc: 0.4141 (0.4158)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.4899 (2.0233)
Test:  [ 0/50]  eta: 0:11:25  loss: 2.2735 (2.2735)  acc1: 48.0000 (48.0000)  acc5: 79.2000 (79.2000)  time: 13.7045  data: 13.6744  max mem: 2905
Test:  [10/50]  eta: 0:01:30  loss: 2.2915 (2.2470)  acc1: 50.4000 (50.5455)  acc5: 76.8000 (76.2182)  time: 2.2507  data: 2.2287  max mem: 2905
Test:  [20/50]  eta: 0:00:49  loss: 2.3651 (2.3817)  acc1: 46.4000 (47.6952)  acc5: 74.4000 (74.5905)  time: 1.0518  data: 1.0316  max mem: 2905
Test:  [30/50]  eta: 0:00:28  loss: 2.4947 (2.3762)  acc1: 44.8000 (47.8710)  acc5: 72.8000 (74.2194)  time: 0.9471  data: 0.9277  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 2.4947 (2.4134)  acc1: 48.0000 (47.4927)  acc5: 70.4000 (73.0341)  time: 0.6323  data: 0.6124  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 2.4857 (2.4287)  acc1: 47.2000 (47.3760)  acc5: 68.8000 (72.4640)  time: 0.6758  data: 0.6558  max mem: 2905
Test: Total time: 0:00:49 (0.9914 s / it)
* Acc@1 47.818 Acc@5 72.532 loss 2.410
Accuracy of the model on the 50000 test images: 47.8%
Max accuracy: 50.54%
Epoch: [32]  [  0/625]  eta: 3:29:47  lr: 0.003982  min_lr: 0.003982  loss: 3.4008 (3.4008)  class_acc: 0.4297 (0.4297)  weight_decay: 0.0500 (0.0500)  time: 20.1404  data: 16.2818  max mem: 2905
Epoch: [32]  [200/625]  eta: 0:14:22  lr: 0.003981  min_lr: 0.003981  loss: 3.5179 (3.4990)  class_acc: 0.4180 (0.4237)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2271 (2.1251)  time: 1.7957  data: 0.0008  max mem: 2905
Epoch: [32]  [400/625]  eta: 0:07:14  lr: 0.003980  min_lr: 0.003980  loss: 3.4666 (3.5082)  class_acc: 0.4336 (0.4210)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2451 (2.2061)  time: 1.9811  data: 0.0006  max mem: 2905
Epoch: [32]  [600/625]  eta: 0:00:48  lr: 0.003979  min_lr: 0.003979  loss: 3.4920 (3.5121)  class_acc: 0.4102 (0.4188)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.6433 (2.0921)  time: 1.8876  data: 0.0009  max mem: 2905
Epoch: [32]  [624/625]  eta: 0:00:01  lr: 0.003979  min_lr: 0.003979  loss: 3.4904 (3.5116)  class_acc: 0.4062 (0.4187)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.6433 (2.0771)  time: 0.7863  data: 0.0018  max mem: 2905
Epoch: [32] Total time: 0:19:50 (1.9042 s / it)
Averaged stats: lr: 0.003979  min_lr: 0.003979  loss: 3.4904 (3.5114)  class_acc: 0.4062 (0.4186)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.6433 (2.0771)
Test:  [ 0/50]  eta: 0:10:48  loss: 2.6732 (2.6732)  acc1: 39.2000 (39.2000)  acc5: 68.0000 (68.0000)  time: 12.9644  data: 12.9375  max mem: 2905
Test:  [10/50]  eta: 0:01:23  loss: 2.3532 (2.4733)  acc1: 48.0000 (46.9091)  acc5: 73.6000 (71.7818)  time: 2.0976  data: 2.0761  max mem: 2905
Test:  [20/50]  eta: 0:00:47  loss: 2.6492 (2.6222)  acc1: 42.4000 (43.5429)  acc5: 70.4000 (69.8667)  time: 1.0304  data: 1.0106  max mem: 2905
Test:  [30/50]  eta: 0:00:27  loss: 2.7581 (2.6196)  acc1: 41.6000 (43.9226)  acc5: 68.0000 (69.3936)  time: 1.0148  data: 0.9960  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 2.7418 (2.6598)  acc1: 41.6000 (43.5317)  acc5: 68.0000 (68.6634)  time: 0.7184  data: 0.6994  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 2.6973 (2.6637)  acc1: 42.4000 (43.6800)  acc5: 68.0000 (68.6720)  time: 0.6728  data: 0.6523  max mem: 2905
Test: Total time: 0:00:48 (0.9736 s / it)
* Acc@1 43.536 Acc@5 68.598 loss 2.643
Accuracy of the model on the 50000 test images: 43.5%
Max accuracy: 50.54%
Epoch: [33]  [  0/625]  eta: 3:21:09  lr: 0.003979  min_lr: 0.003979  loss: 3.6944 (3.6944)  class_acc: 0.3672 (0.3672)  weight_decay: 0.0500 (0.0500)  time: 19.3115  data: 16.1847  max mem: 2905
Epoch: [33]  [200/625]  eta: 0:13:48  lr: 0.003978  min_lr: 0.003978  loss: 3.4921 (3.4781)  class_acc: 0.4180 (0.4240)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.6415 (2.2971)  time: 1.7956  data: 0.0209  max mem: 2905
Epoch: [33]  [400/625]  eta: 0:07:05  lr: 0.003977  min_lr: 0.003977  loss: 3.4732 (3.4920)  class_acc: 0.4297 (0.4222)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.5325 (2.2462)  time: 1.9609  data: 0.0069  max mem: 2905
Epoch: [33]  [600/625]  eta: 0:00:48  lr: 0.003976  min_lr: 0.003976  loss: 3.4680 (3.4957)  class_acc: 0.4219 (0.4212)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.4542 (2.2352)  time: 2.1684  data: 0.0011  max mem: 2905
Epoch: [33]  [624/625]  eta: 0:00:01  lr: 0.003975  min_lr: 0.003975  loss: 3.5286 (3.4964)  class_acc: 0.4141 (0.4210)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.5404 (2.2102)  time: 0.6252  data: 0.0030  max mem: 2905
Epoch: [33] Total time: 0:19:59 (1.9199 s / it)
Averaged stats: lr: 0.003975  min_lr: 0.003975  loss: 3.5286 (3.5017)  class_acc: 0.4141 (0.4206)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.5404 (2.2102)
Test:  [ 0/50]  eta: 0:11:32  loss: 2.2907 (2.2907)  acc1: 50.4000 (50.4000)  acc5: 75.2000 (75.2000)  time: 13.8549  data: 13.8182  max mem: 2905
Test:  [10/50]  eta: 0:01:38  loss: 2.2598 (2.2844)  acc1: 50.4000 (50.9818)  acc5: 75.2000 (74.4000)  time: 2.4684  data: 2.4484  max mem: 2905
Test:  [20/50]  eta: 0:00:56  loss: 2.3548 (2.3906)  acc1: 48.0000 (48.9143)  acc5: 75.2000 (73.6000)  time: 1.2919  data: 1.2731  max mem: 2905
Test:  [30/50]  eta: 0:00:31  loss: 2.4564 (2.3973)  acc1: 46.4000 (48.0516)  acc5: 72.8000 (73.3419)  time: 1.0758  data: 1.0563  max mem: 2905
Test:  [40/50]  eta: 0:00:13  loss: 2.5022 (2.4034)  acc1: 46.4000 (47.9610)  acc5: 72.0000 (73.3463)  time: 0.6888  data: 0.6695  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.4441 (2.4173)  acc1: 46.4000 (47.7440)  acc5: 71.2000 (73.1360)  time: 0.6201  data: 0.6010  max mem: 2905
Test: Total time: 0:00:54 (1.0916 s / it)
* Acc@1 48.582 Acc@5 73.274 loss 2.370
Accuracy of the model on the 50000 test images: 48.6%
Max accuracy: 50.54%
Epoch: [34]  [  0/625]  eta: 3:23:05  lr: 0.003975  min_lr: 0.003975  loss: 3.4742 (3.4742)  class_acc: 0.4258 (0.4258)  weight_decay: 0.0500 (0.0500)  time: 19.4971  data: 17.0031  max mem: 2905
Epoch: [34]  [200/625]  eta: 0:14:41  lr: 0.003974  min_lr: 0.003974  loss: 3.4882 (3.4755)  class_acc: 0.4258 (0.4260)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.2058 (2.2156)  time: 1.8631  data: 0.0006  max mem: 2905
Epoch: [34]  [400/625]  eta: 0:07:33  lr: 0.003973  min_lr: 0.003973  loss: 3.4412 (3.4874)  class_acc: 0.4258 (0.4244)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0687 (inf)  time: 1.9048  data: 0.0007  max mem: 2905
Epoch: [34]  [600/625]  eta: 0:00:50  lr: 0.003972  min_lr: 0.003972  loss: 3.4508 (3.4869)  class_acc: 0.4219 (0.4240)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.4027 (inf)  time: 1.8883  data: 0.0006  max mem: 2905
Epoch: [34]  [624/625]  eta: 0:00:01  lr: 0.003972  min_lr: 0.003972  loss: 3.4583 (3.4864)  class_acc: 0.4219 (0.4241)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.6445 (inf)  time: 0.8677  data: 0.0014  max mem: 2905
Epoch: [34] Total time: 0:20:24 (1.9595 s / it)
Averaged stats: lr: 0.003972  min_lr: 0.003972  loss: 3.4583 (3.4904)  class_acc: 0.4219 (0.4227)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.6445 (inf)
Test:  [ 0/50]  eta: 0:10:56  loss: 2.3697 (2.3697)  acc1: 48.0000 (48.0000)  acc5: 74.4000 (74.4000)  time: 13.1339  data: 13.1075  max mem: 2905
Test:  [10/50]  eta: 0:01:26  loss: 2.3371 (2.3366)  acc1: 50.4000 (50.5455)  acc5: 73.6000 (74.1091)  time: 2.1663  data: 2.1474  max mem: 2905
Test:  [20/50]  eta: 0:00:50  loss: 2.4313 (2.4743)  acc1: 47.2000 (47.0095)  acc5: 71.2000 (72.2667)  time: 1.1212  data: 1.1027  max mem: 2905
Test:  [30/50]  eta: 0:00:29  loss: 2.6178 (2.4762)  acc1: 44.8000 (46.6323)  acc5: 70.4000 (71.7936)  time: 1.0683  data: 1.0487  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 2.5198 (2.4820)  acc1: 46.4000 (46.6927)  acc5: 71.2000 (71.6488)  time: 0.6772  data: 0.6585  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 2.4586 (2.4786)  acc1: 45.6000 (46.7840)  acc5: 72.0000 (71.6960)  time: 0.5905  data: 0.5721  max mem: 2905
Test: Total time: 0:00:50 (1.0058 s / it)
* Acc@1 47.320 Acc@5 72.012 loss 2.447
Accuracy of the model on the 50000 test images: 47.3%
Max accuracy: 50.54%
Epoch: [35]  [  0/625]  eta: 3:39:56  lr: 0.003972  min_lr: 0.003972  loss: 3.3530 (3.3530)  class_acc: 0.4297 (0.4297)  weight_decay: 0.0500 (0.0500)  time: 21.1136  data: 16.9161  max mem: 2905
Epoch: [35]  [200/625]  eta: 0:14:37  lr: 0.003971  min_lr: 0.003971  loss: 3.4848 (3.4611)  class_acc: 0.4141 (0.4300)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7237 (2.2412)  time: 1.7407  data: 0.0104  max mem: 2905
Epoch: [35]  [400/625]  eta: 0:07:26  lr: 0.003969  min_lr: 0.003969  loss: 3.4997 (3.4627)  class_acc: 0.4258 (0.4276)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.6701 (2.2118)  time: 1.8030  data: 0.0653  max mem: 2905
Epoch: [35]  [600/625]  eta: 0:00:49  lr: 0.003968  min_lr: 0.003968  loss: 3.5058 (3.4728)  class_acc: 0.4219 (0.4254)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.8002 (2.2541)  time: 2.0367  data: 0.0062  max mem: 2905
Epoch: [35]  [624/625]  eta: 0:00:01  lr: 0.003968  min_lr: 0.003968  loss: 3.5165 (3.4741)  class_acc: 0.4219 (0.4252)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8629 (2.2401)  time: 0.4576  data: 0.0141  max mem: 2905
Epoch: [35] Total time: 0:20:22 (1.9558 s / it)
Averaged stats: lr: 0.003968  min_lr: 0.003968  loss: 3.5165 (3.4797)  class_acc: 0.4219 (0.4247)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8629 (2.2401)
Test:  [ 0/50]  eta: 0:10:42  loss: 2.6505 (2.6505)  acc1: 37.6000 (37.6000)  acc5: 69.6000 (69.6000)  time: 12.8504  data: 12.8258  max mem: 2905
Test:  [10/50]  eta: 0:01:25  loss: 2.4185 (2.3711)  acc1: 52.0000 (49.5273)  acc5: 74.4000 (73.6000)  time: 2.1436  data: 2.1250  max mem: 2905
Test:  [20/50]  eta: 0:00:49  loss: 2.4582 (2.4916)  acc1: 45.6000 (46.1333)  acc5: 72.8000 (71.6571)  time: 1.0952  data: 1.0773  max mem: 2905
Test:  [30/50]  eta: 0:00:29  loss: 2.6083 (2.5177)  acc1: 42.4000 (45.5484)  acc5: 69.6000 (71.4581)  time: 1.0996  data: 1.0820  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 2.6142 (2.5576)  acc1: 44.0000 (45.3659)  acc5: 68.0000 (70.4585)  time: 0.8511  data: 0.8329  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.6923 (2.5644)  acc1: 42.4000 (44.9280)  acc5: 68.0000 (70.2240)  time: 0.7552  data: 0.7365  max mem: 2905
Test: Total time: 0:00:54 (1.0944 s / it)
* Acc@1 45.012 Acc@5 70.206 loss 2.546
Accuracy of the model on the 50000 test images: 45.0%
Max accuracy: 50.54%
Epoch: [36]  [  0/625]  eta: 3:58:20  lr: 0.003968  min_lr: 0.003968  loss: 3.3311 (3.3311)  class_acc: 0.4453 (0.4453)  weight_decay: 0.0500 (0.0500)  time: 22.8804  data: 18.0996  max mem: 2905
Epoch: [36]  [200/625]  eta: 0:14:25  lr: 0.003967  min_lr: 0.003967  loss: 3.4543 (3.4543)  class_acc: 0.4336 (0.4306)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4093 (2.3944)  time: 2.0214  data: 0.0008  max mem: 2905
Epoch: [36]  [400/625]  eta: 0:07:26  lr: 0.003965  min_lr: 0.003965  loss: 3.4779 (3.4624)  class_acc: 0.4180 (0.4283)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.6931 (2.1741)  time: 1.9241  data: 0.0102  max mem: 2905
Epoch: [36]  [600/625]  eta: 0:00:49  lr: 0.003964  min_lr: 0.003964  loss: 3.3859 (3.4684)  class_acc: 0.4297 (0.4278)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.4834 (2.1673)  time: 2.1073  data: 0.0040  max mem: 2905
Epoch: [36]  [624/625]  eta: 0:00:01  lr: 0.003964  min_lr: 0.003964  loss: 3.4411 (3.4672)  class_acc: 0.4219 (0.4281)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7825 (2.1718)  time: 0.6709  data: 0.0014  max mem: 2905
Epoch: [36] Total time: 0:20:11 (1.9380 s / it)
Averaged stats: lr: 0.003964  min_lr: 0.003964  loss: 3.4411 (3.4641)  class_acc: 0.4219 (0.4281)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7825 (2.1718)
Test:  [ 0/50]  eta: 0:08:49  loss: 2.0329 (2.0329)  acc1: 52.8000 (52.8000)  acc5: 80.0000 (80.0000)  time: 10.5943  data: 10.5649  max mem: 2905
Test:  [10/50]  eta: 0:01:08  loss: 2.1859 (2.2179)  acc1: 52.8000 (53.0909)  acc5: 76.0000 (75.0545)  time: 1.7098  data: 1.6906  max mem: 2905
Test:  [20/50]  eta: 0:00:42  loss: 2.3918 (2.3210)  acc1: 50.4000 (49.4095)  acc5: 72.8000 (73.6381)  time: 0.9658  data: 0.9471  max mem: 2905
Test:  [30/50]  eta: 0:00:27  loss: 2.4170 (2.3351)  acc1: 46.4000 (49.1871)  acc5: 71.2000 (73.5484)  time: 1.1629  data: 1.1430  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 2.3483 (2.3676)  acc1: 47.2000 (48.5463)  acc5: 71.2000 (73.1122)  time: 0.9476  data: 0.9282  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.4027 (2.3716)  acc1: 45.6000 (48.4640)  acc5: 73.6000 (73.1520)  time: 0.5007  data: 0.4828  max mem: 2905
Test: Total time: 0:00:52 (1.0439 s / it)
* Acc@1 49.020 Acc@5 73.458 loss 2.348
Accuracy of the model on the 50000 test images: 49.0%
Max accuracy: 50.54%
Epoch: [37]  [  0/625]  eta: 3:44:21  lr: 0.003964  min_lr: 0.003964  loss: 3.5528 (3.5528)  class_acc: 0.4375 (0.4375)  weight_decay: 0.0500 (0.0500)  time: 21.5379  data: 21.2904  max mem: 2905
Epoch: [37]  [200/625]  eta: 0:14:39  lr: 0.003962  min_lr: 0.003962  loss: 3.4587 (3.4448)  class_acc: 0.4141 (0.4326)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9882 (2.3421)  time: 2.0363  data: 0.2053  max mem: 2905
Epoch: [37]  [400/625]  eta: 0:07:31  lr: 0.003961  min_lr: 0.003961  loss: 3.4508 (3.4458)  class_acc: 0.4453 (0.4317)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8675 (2.2136)  time: 1.9872  data: 0.0010  max mem: 2905
Epoch: [37]  [600/625]  eta: 0:00:49  lr: 0.003960  min_lr: 0.003960  loss: 3.4958 (3.4564)  class_acc: 0.4141 (0.4294)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.3652 (2.1520)  time: 2.1020  data: 0.0366  max mem: 2905
Epoch: [37]  [624/625]  eta: 0:00:01  lr: 0.003959  min_lr: 0.003959  loss: 3.5078 (3.4588)  class_acc: 0.4180 (0.4289)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.3907 (2.1503)  time: 0.6706  data: 0.0018  max mem: 2905
Epoch: [37] Total time: 0:20:12 (1.9398 s / it)
Averaged stats: lr: 0.003959  min_lr: 0.003959  loss: 3.5078 (3.4561)  class_acc: 0.4180 (0.4296)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.3907 (2.1503)
Test:  [ 0/50]  eta: 0:09:47  loss: 2.3570 (2.3570)  acc1: 48.8000 (48.8000)  acc5: 76.0000 (76.0000)  time: 11.7520  data: 11.7171  max mem: 2905
Test:  [10/50]  eta: 0:01:15  loss: 2.1305 (2.1705)  acc1: 52.8000 (51.7091)  acc5: 77.6000 (77.0182)  time: 1.8980  data: 1.8762  max mem: 2905
Test:  [20/50]  eta: 0:00:43  loss: 2.3069 (2.3413)  acc1: 47.2000 (48.5714)  acc5: 74.4000 (74.5143)  time: 0.9281  data: 0.9079  max mem: 2905
Test:  [30/50]  eta: 0:00:24  loss: 2.5020 (2.3560)  acc1: 44.8000 (47.8968)  acc5: 72.0000 (73.8065)  time: 0.8678  data: 0.8476  max mem: 2905
Test:  [40/50]  eta: 0:00:10  loss: 2.4412 (2.3812)  acc1: 45.6000 (47.9415)  acc5: 72.0000 (73.3463)  time: 0.6246  data: 0.6049  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 2.3646 (2.3709)  acc1: 47.2000 (47.9840)  acc5: 72.0000 (73.5040)  time: 0.4498  data: 0.4308  max mem: 2905
Test: Total time: 0:00:45 (0.9003 s / it)
* Acc@1 49.484 Acc@5 74.074 loss 2.317
Accuracy of the model on the 50000 test images: 49.5%
Max accuracy: 50.54%
Epoch: [38]  [  0/625]  eta: 3:34:39  lr: 0.003959  min_lr: 0.003959  loss: 3.5082 (3.5082)  class_acc: 0.4180 (0.4180)  weight_decay: 0.0500 (0.0500)  time: 20.6079  data: 20.1486  max mem: 2905
Epoch: [38]  [200/625]  eta: 0:14:18  lr: 0.003958  min_lr: 0.003958  loss: 3.4614 (3.4361)  class_acc: 0.4258 (0.4308)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.6697 (2.3046)  time: 1.9057  data: 0.0942  max mem: 2905
Epoch: [38]  [400/625]  eta: 0:07:27  lr: 0.003956  min_lr: 0.003956  loss: 3.4751 (3.4424)  class_acc: 0.4141 (0.4296)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.4432 (2.4068)  time: 1.9917  data: 0.0008  max mem: 2905
Epoch: [38]  [600/625]  eta: 0:00:49  lr: 0.003955  min_lr: 0.003955  loss: 3.4399 (3.4434)  class_acc: 0.4297 (0.4302)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7011 (2.3015)  time: 2.2151  data: 0.0007  max mem: 2905
Epoch: [38]  [624/625]  eta: 0:00:01  lr: 0.003955  min_lr: 0.003955  loss: 3.4671 (3.4436)  class_acc: 0.4219 (0.4301)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9491 (2.3007)  time: 0.6622  data: 0.0013  max mem: 2905
Epoch: [38] Total time: 0:20:12 (1.9395 s / it)
Averaged stats: lr: 0.003955  min_lr: 0.003955  loss: 3.4671 (3.4478)  class_acc: 0.4219 (0.4311)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9491 (2.3007)
Test:  [ 0/50]  eta: 0:10:14  loss: 2.7064 (2.7064)  acc1: 40.0000 (40.0000)  acc5: 72.0000 (72.0000)  time: 12.2815  data: 12.2534  max mem: 2905
Test:  [10/50]  eta: 0:01:25  loss: 2.5738 (2.5380)  acc1: 45.6000 (45.0909)  acc5: 71.2000 (70.9818)  time: 2.1405  data: 2.1216  max mem: 2905
Test:  [20/50]  eta: 0:00:52  loss: 2.7629 (2.6967)  acc1: 43.2000 (42.4762)  acc5: 68.8000 (69.1048)  time: 1.2291  data: 1.2106  max mem: 2905
Test:  [30/50]  eta: 0:00:31  loss: 2.8041 (2.6914)  acc1: 40.0000 (42.4000)  acc5: 67.2000 (68.5677)  time: 1.2787  data: 1.2603  max mem: 2905
Test:  [40/50]  eta: 0:00:13  loss: 2.6971 (2.7138)  acc1: 41.6000 (42.1073)  acc5: 68.0000 (68.0390)  time: 0.8855  data: 0.8666  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.7160 (2.7101)  acc1: 43.2000 (42.5280)  acc5: 68.0000 (68.0480)  time: 0.7354  data: 0.7161  max mem: 2905
Test: Total time: 0:00:55 (1.1124 s / it)
* Acc@1 43.506 Acc@5 68.474 loss 2.667
Accuracy of the model on the 50000 test images: 43.5%
Max accuracy: 50.54%
Epoch: [39]  [  0/625]  eta: 3:45:22  lr: 0.003955  min_lr: 0.003955  loss: 3.1841 (3.1841)  class_acc: 0.4766 (0.4766)  weight_decay: 0.0500 (0.0500)  time: 21.6368  data: 18.6295  max mem: 2905
Epoch: [39]  [200/625]  eta: 0:15:01  lr: 0.003953  min_lr: 0.003953  loss: 3.4117 (3.4293)  class_acc: 0.4336 (0.4337)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.4930 (2.4655)  time: 1.9263  data: 0.0006  max mem: 2905
Epoch: [39]  [400/625]  eta: 0:07:42  lr: 0.003952  min_lr: 0.003952  loss: 3.3971 (3.4350)  class_acc: 0.4297 (0.4320)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.6770 (2.2379)  time: 1.9215  data: 0.0007  max mem: 2905
Epoch: [39]  [600/625]  eta: 0:00:51  lr: 0.003950  min_lr: 0.003950  loss: 3.4286 (3.4383)  class_acc: 0.4414 (0.4327)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7661 (2.1673)  time: 2.0051  data: 0.0007  max mem: 2905
Epoch: [39]  [624/625]  eta: 0:00:01  lr: 0.003950  min_lr: 0.003950  loss: 3.4282 (3.4385)  class_acc: 0.4297 (0.4327)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.5666 (2.1542)  time: 0.7318  data: 0.0014  max mem: 2905
Epoch: [39] Total time: 0:20:43 (1.9903 s / it)
Averaged stats: lr: 0.003950  min_lr: 0.003950  loss: 3.4282 (3.4390)  class_acc: 0.4297 (0.4330)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.5666 (2.1542)
Test:  [ 0/50]  eta: 0:10:55  loss: 2.3686 (2.3686)  acc1: 44.8000 (44.8000)  acc5: 72.8000 (72.8000)  time: 13.1138  data: 13.0878  max mem: 2905
Test:  [10/50]  eta: 0:01:23  loss: 2.1979 (2.1789)  acc1: 53.6000 (53.7455)  acc5: 75.2000 (75.7091)  time: 2.0881  data: 2.0680  max mem: 2905
Test:  [20/50]  eta: 0:00:48  loss: 2.3461 (2.3167)  acc1: 49.6000 (50.3619)  acc5: 73.6000 (74.0191)  time: 1.0489  data: 1.0295  max mem: 2905
Test:  [30/50]  eta: 0:00:28  loss: 2.3705 (2.3122)  acc1: 48.0000 (50.3742)  acc5: 73.6000 (73.9355)  time: 1.0630  data: 1.0444  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 2.3351 (2.3320)  acc1: 48.8000 (49.8537)  acc5: 73.6000 (73.5220)  time: 0.7608  data: 0.7425  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.3770 (2.3480)  acc1: 47.2000 (49.2960)  acc5: 72.8000 (73.3920)  time: 0.6592  data: 0.6413  max mem: 2905
Test: Total time: 0:00:50 (1.0056 s / it)
* Acc@1 49.532 Acc@5 73.846 loss 2.324
Accuracy of the model on the 50000 test images: 49.5%
Max accuracy: 50.54%
Epoch: [40]  [  0/625]  eta: 3:28:09  lr: 0.003950  min_lr: 0.003950  loss: 3.5175 (3.5175)  class_acc: 0.4023 (0.4023)  weight_decay: 0.0500 (0.0500)  time: 19.9833  data: 15.8166  max mem: 2905
Epoch: [40]  [200/625]  eta: 0:13:59  lr: 0.003948  min_lr: 0.003948  loss: 3.4259 (3.4236)  class_acc: 0.4297 (0.4370)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3682 (2.3779)  time: 1.9172  data: 0.2894  max mem: 2905
Epoch: [40]  [400/625]  eta: 0:07:14  lr: 0.003947  min_lr: 0.003947  loss: 3.3919 (3.4230)  class_acc: 0.4297 (0.4370)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8893 (2.2478)  time: 1.7788  data: 0.0706  max mem: 2905
Epoch: [40]  [600/625]  eta: 0:00:47  lr: 0.003945  min_lr: 0.003945  loss: 3.4431 (3.4283)  class_acc: 0.4336 (0.4363)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.6500 (2.1530)  time: 1.8195  data: 0.2604  max mem: 2905
Epoch: [40]  [624/625]  eta: 0:00:01  lr: 0.003945  min_lr: 0.003945  loss: 3.4649 (3.4298)  class_acc: 0.4297 (0.4360)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.6349 (2.1598)  time: 0.7839  data: 0.1111  max mem: 2905
Epoch: [40] Total time: 0:19:32 (1.8761 s / it)
Averaged stats: lr: 0.003945  min_lr: 0.003945  loss: 3.4649 (3.4282)  class_acc: 0.4297 (0.4356)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.6349 (2.1598)
Test:  [ 0/50]  eta: 0:09:56  loss: 2.6887 (2.6887)  acc1: 40.8000 (40.8000)  acc5: 68.0000 (68.0000)  time: 11.9333  data: 11.9080  max mem: 2905
Test:  [10/50]  eta: 0:01:20  loss: 2.4430 (2.3694)  acc1: 48.8000 (50.3273)  acc5: 72.8000 (73.4545)  time: 2.0103  data: 1.9918  max mem: 2905
Test:  [20/50]  eta: 0:00:46  loss: 2.6287 (2.5685)  acc1: 44.8000 (45.6000)  acc5: 68.0000 (70.3238)  time: 1.0345  data: 1.0155  max mem: 2905
Test:  [30/50]  eta: 0:00:26  loss: 2.6538 (2.5495)  acc1: 42.4000 (45.5484)  acc5: 68.0000 (70.5806)  time: 0.9829  data: 0.9632  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 2.5278 (2.5705)  acc1: 43.2000 (45.4634)  acc5: 68.8000 (69.9317)  time: 0.7307  data: 0.7109  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 2.5604 (2.5863)  acc1: 44.8000 (45.1360)  acc5: 68.8000 (69.7600)  time: 0.6761  data: 0.6564  max mem: 2905
Test: Total time: 0:00:48 (0.9642 s / it)
* Acc@1 45.532 Acc@5 70.668 loss 2.553
Accuracy of the model on the 50000 test images: 45.5%
Max accuracy: 50.54%
Epoch: [41]  [  0/625]  eta: 3:26:26  lr: 0.003945  min_lr: 0.003945  loss: 3.5007 (3.5007)  class_acc: 0.4336 (0.4336)  weight_decay: 0.0500 (0.0500)  time: 19.8185  data: 17.3438  max mem: 2905
Epoch: [41]  [200/625]  eta: 0:14:00  lr: 0.003943  min_lr: 0.003943  loss: 3.3732 (3.3916)  class_acc: 0.4375 (0.4430)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8629 (2.4691)  time: 1.8012  data: 0.3605  max mem: 2905
Epoch: [41]  [400/625]  eta: 0:07:15  lr: 0.003941  min_lr: 0.003941  loss: 3.4659 (3.4016)  class_acc: 0.4219 (0.4415)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.4757 (inf)  time: 1.9578  data: 0.0576  max mem: 2905
Epoch: [41]  [600/625]  eta: 0:00:48  lr: 0.003940  min_lr: 0.003940  loss: 3.4806 (3.4165)  class_acc: 0.4258 (0.4386)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9142 (inf)  time: 1.9242  data: 0.0218  max mem: 2905
Epoch: [41]  [624/625]  eta: 0:00:01  lr: 0.003939  min_lr: 0.003939  loss: 3.3626 (3.4157)  class_acc: 0.4492 (0.4387)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.6419 (inf)  time: 0.9451  data: 0.0015  max mem: 2905
Epoch: [41] Total time: 0:19:49 (1.9028 s / it)
Averaged stats: lr: 0.003939  min_lr: 0.003939  loss: 3.3626 (3.4211)  class_acc: 0.4492 (0.4371)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.6419 (inf)
Test:  [ 0/50]  eta: 0:10:34  loss: 2.2427 (2.2427)  acc1: 50.4000 (50.4000)  acc5: 77.6000 (77.6000)  time: 12.6959  data: 12.6652  max mem: 2905
Test:  [10/50]  eta: 0:01:27  loss: 2.2100 (2.2620)  acc1: 52.8000 (51.9273)  acc5: 76.0000 (75.6364)  time: 2.1897  data: 2.1676  max mem: 2905
Test:  [20/50]  eta: 0:00:51  loss: 2.4331 (2.3792)  acc1: 47.2000 (48.4571)  acc5: 74.4000 (74.2857)  time: 1.1603  data: 1.1400  max mem: 2905
Test:  [30/50]  eta: 0:00:29  loss: 2.4820 (2.4063)  acc1: 44.8000 (47.7936)  acc5: 72.8000 (73.8323)  time: 1.0871  data: 1.0658  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 2.4954 (2.4318)  acc1: 44.8000 (47.4537)  acc5: 71.2000 (73.1122)  time: 0.6651  data: 0.6445  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 2.4891 (2.4456)  acc1: 44.8000 (47.3920)  acc5: 71.2000 (72.9280)  time: 0.6625  data: 0.6444  max mem: 2905
Test: Total time: 0:00:49 (0.9910 s / it)
* Acc@1 47.910 Acc@5 72.916 loss 2.415
Accuracy of the model on the 50000 test images: 47.9%
Max accuracy: 50.54%
Epoch: [42]  [  0/625]  eta: 3:33:02  lr: 0.003939  min_lr: 0.003939  loss: 3.1996 (3.1996)  class_acc: 0.5078 (0.5078)  weight_decay: 0.0500 (0.0500)  time: 20.4526  data: 17.3140  max mem: 2905
Epoch: [42]  [200/625]  eta: 0:13:47  lr: 0.003938  min_lr: 0.003938  loss: 3.3309 (3.3978)  class_acc: 0.4531 (0.4408)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2845 (2.5679)  time: 1.8660  data: 0.0112  max mem: 2905
Epoch: [42]  [400/625]  eta: 0:07:21  lr: 0.003936  min_lr: 0.003936  loss: 3.4166 (3.4028)  class_acc: 0.4375 (0.4404)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.4697 (2.3329)  time: 2.1669  data: 0.0209  max mem: 2905
Epoch: [42]  [600/625]  eta: 0:00:50  lr: 0.003934  min_lr: 0.003934  loss: 3.3962 (3.4140)  class_acc: 0.4453 (0.4388)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1224 (2.3716)  time: 2.0179  data: 0.0056  max mem: 2905
Epoch: [42]  [624/625]  eta: 0:00:02  lr: 0.003934  min_lr: 0.003934  loss: 3.4417 (3.4153)  class_acc: 0.4375 (0.4387)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.6972 (2.3574)  time: 0.7335  data: 0.0014  max mem: 2905
Epoch: [42] Total time: 0:21:02 (2.0197 s / it)
Averaged stats: lr: 0.003934  min_lr: 0.003934  loss: 3.4417 (3.4156)  class_acc: 0.4375 (0.4376)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.6972 (2.3574)
Test:  [ 0/50]  eta: 0:09:13  loss: 2.4724 (2.4724)  acc1: 51.2000 (51.2000)  acc5: 72.0000 (72.0000)  time: 11.0760  data: 11.0492  max mem: 2905
Test:  [10/50]  eta: 0:01:31  loss: 2.3714 (2.3949)  acc1: 48.8000 (48.5091)  acc5: 73.6000 (72.9455)  time: 2.2761  data: 2.2572  max mem: 2905
Test:  [20/50]  eta: 0:00:53  loss: 2.4774 (2.5129)  acc1: 46.4000 (45.5238)  acc5: 72.8000 (71.4667)  time: 1.3361  data: 1.3179  max mem: 2905
Test:  [30/50]  eta: 0:00:30  loss: 2.5873 (2.5127)  acc1: 44.8000 (45.6774)  acc5: 70.4000 (71.4323)  time: 1.1018  data: 1.0825  max mem: 2905
Test:  [40/50]  eta: 0:00:13  loss: 2.5052 (2.5162)  acc1: 46.4000 (45.8146)  acc5: 72.0000 (71.4732)  time: 0.7906  data: 0.7713  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.4843 (2.5266)  acc1: 45.6000 (45.9840)  acc5: 72.0000 (71.4080)  time: 0.7060  data: 0.6867  max mem: 2905
Test: Total time: 0:00:55 (1.1005 s / it)
* Acc@1 46.254 Acc@5 71.254 loss 2.505
Accuracy of the model on the 50000 test images: 46.3%
Max accuracy: 50.54%
Epoch: [43]  [  0/625]  eta: 3:39:59  lr: 0.003934  min_lr: 0.003934  loss: 3.3257 (3.3257)  class_acc: 0.4531 (0.4531)  weight_decay: 0.0500 (0.0500)  time: 21.1192  data: 19.2549  max mem: 2905
Epoch: [43]  [200/625]  eta: 0:14:44  lr: 0.003932  min_lr: 0.003932  loss: 3.4162 (3.4085)  class_acc: 0.4414 (0.4399)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7694 (2.3014)  time: 1.9873  data: 0.0009  max mem: 2905
Epoch: [43]  [400/625]  eta: 0:07:39  lr: 0.003930  min_lr: 0.003930  loss: 3.3546 (3.4110)  class_acc: 0.4570 (0.4383)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.6722 (2.2402)  time: 2.1345  data: 0.0006  max mem: 2905
Epoch: [43]  [600/625]  eta: 0:00:51  lr: 0.003928  min_lr: 0.003928  loss: 3.4291 (3.4132)  class_acc: 0.4258 (0.4374)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7748 (2.2886)  time: 1.9398  data: 0.0007  max mem: 2905
Epoch: [43]  [624/625]  eta: 0:00:02  lr: 0.003928  min_lr: 0.003928  loss: 3.3666 (3.4122)  class_acc: 0.4531 (0.4376)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0819 (2.3060)  time: 0.8716  data: 0.0013  max mem: 2905
Epoch: [43] Total time: 0:20:51 (2.0021 s / it)
Averaged stats: lr: 0.003928  min_lr: 0.003928  loss: 3.3666 (3.4030)  class_acc: 0.4531 (0.4402)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0819 (2.3060)
Test:  [ 0/50]  eta: 0:10:58  loss: 2.6287 (2.6287)  acc1: 46.4000 (46.4000)  acc5: 72.0000 (72.0000)  time: 13.1616  data: 13.1340  max mem: 2905
Test:  [10/50]  eta: 0:01:25  loss: 2.2934 (2.3228)  acc1: 49.6000 (50.2545)  acc5: 74.4000 (73.8909)  time: 2.1315  data: 2.1103  max mem: 2905
Test:  [20/50]  eta: 0:00:49  loss: 2.4850 (2.4914)  acc1: 47.2000 (46.6286)  acc5: 72.8000 (72.4952)  time: 1.0617  data: 1.0402  max mem: 2905
Test:  [30/50]  eta: 0:00:28  loss: 2.6065 (2.4646)  acc1: 42.4000 (46.7355)  acc5: 71.2000 (72.4129)  time: 1.0437  data: 1.0223  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 2.5615 (2.4729)  acc1: 44.0000 (46.6732)  acc5: 70.4000 (72.3122)  time: 0.8354  data: 0.8146  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.5612 (2.4825)  acc1: 45.6000 (46.5920)  acc5: 70.4000 (72.2080)  time: 0.7857  data: 0.7649  max mem: 2905
Test: Total time: 0:00:55 (1.1032 s / it)
* Acc@1 47.476 Acc@5 72.522 loss 2.442
Accuracy of the model on the 50000 test images: 47.5%
Max accuracy: 50.54%
Epoch: [44]  [  0/625]  eta: 4:18:49  lr: 0.003928  min_lr: 0.003928  loss: 3.1990 (3.1990)  class_acc: 0.4922 (0.4922)  weight_decay: 0.0500 (0.0500)  time: 24.8479  data: 22.2422  max mem: 2905
Epoch: [44]  [200/625]  eta: 0:15:02  lr: 0.003926  min_lr: 0.003926  loss: 3.4597 (3.3822)  class_acc: 0.4297 (0.4461)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5751 (2.3647)  time: 2.0278  data: 0.0237  max mem: 2905
Epoch: [44]  [400/625]  eta: 0:07:30  lr: 0.003924  min_lr: 0.003924  loss: 3.4122 (3.3933)  class_acc: 0.4453 (0.4438)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.6542 (2.2966)  time: 1.8663  data: 1.6861  max mem: 2905
Epoch: [44]  [600/625]  eta: 0:00:50  lr: 0.003922  min_lr: 0.003922  loss: 3.3890 (3.3998)  class_acc: 0.4375 (0.4425)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.4126 (2.2748)  time: 1.9420  data: 0.7089  max mem: 2905
Epoch: [44]  [624/625]  eta: 0:00:01  lr: 0.003922  min_lr: 0.003922  loss: 3.4104 (3.4009)  class_acc: 0.4453 (0.4426)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5555 (2.3638)  time: 0.8427  data: 0.3045  max mem: 2905
Epoch: [44] Total time: 0:20:24 (1.9587 s / it)
Averaged stats: lr: 0.003922  min_lr: 0.003922  loss: 3.4104 (3.3967)  class_acc: 0.4453 (0.4417)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5555 (2.3638)
Test:  [ 0/50]  eta: 0:09:43  loss: 2.8062 (2.8062)  acc1: 40.8000 (40.8000)  acc5: 68.0000 (68.0000)  time: 11.6620  data: 11.6140  max mem: 2905
Test:  [10/50]  eta: 0:01:23  loss: 2.6173 (2.6161)  acc1: 44.0000 (45.7455)  acc5: 68.8000 (69.4545)  time: 2.0819  data: 2.0603  max mem: 2905
Test:  [20/50]  eta: 0:00:50  loss: 2.7449 (2.7464)  acc1: 40.8000 (42.2476)  acc5: 67.2000 (67.7714)  time: 1.1730  data: 1.1528  max mem: 2905
Test:  [30/50]  eta: 0:00:27  loss: 2.8054 (2.7631)  acc1: 39.2000 (42.1677)  acc5: 65.6000 (67.6903)  time: 0.9937  data: 0.9739  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 2.8054 (2.7753)  acc1: 37.6000 (41.5610)  acc5: 64.8000 (66.9463)  time: 0.5233  data: 0.5035  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 2.7604 (2.7762)  acc1: 41.6000 (41.6160)  acc5: 66.4000 (67.0880)  time: 0.4337  data: 0.4133  max mem: 2905
Test: Total time: 0:00:46 (0.9396 s / it)
* Acc@1 42.200 Acc@5 67.380 loss 2.747
Accuracy of the model on the 50000 test images: 42.2%
Max accuracy: 50.54%
Epoch: [45]  [  0/625]  eta: 3:22:33  lr: 0.003922  min_lr: 0.003922  loss: 3.1638 (3.1638)  class_acc: 0.4766 (0.4766)  weight_decay: 0.0500 (0.0500)  time: 19.4456  data: 16.4242  max mem: 2905
Epoch: [45]  [200/625]  eta: 0:14:08  lr: 0.003920  min_lr: 0.003920  loss: 3.4061 (3.3835)  class_acc: 0.4453 (0.4440)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0137 (2.2698)  time: 1.9818  data: 0.0008  max mem: 2905
Epoch: [45]  [400/625]  eta: 0:07:21  lr: 0.003918  min_lr: 0.003918  loss: 3.3883 (3.3812)  class_acc: 0.4453 (0.4436)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.5594 (2.2185)  time: 2.0136  data: 0.0011  max mem: 2905
Epoch: [45]  [600/625]  eta: 0:00:49  lr: 0.003916  min_lr: 0.003916  loss: 3.3852 (3.3876)  class_acc: 0.4492 (0.4421)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.4303 (2.1894)  time: 2.1442  data: 0.0318  max mem: 2905
Epoch: [45]  [624/625]  eta: 0:00:01  lr: 0.003916  min_lr: 0.003916  loss: 3.4186 (3.3894)  class_acc: 0.4297 (0.4419)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8138 (2.1946)  time: 0.7823  data: 0.0020  max mem: 2905
Epoch: [45] Total time: 0:20:02 (1.9247 s / it)
Averaged stats: lr: 0.003916  min_lr: 0.003916  loss: 3.4186 (3.3921)  class_acc: 0.4297 (0.4426)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8138 (2.1946)
Test:  [ 0/50]  eta: 0:11:25  loss: 2.8380 (2.8380)  acc1: 36.8000 (36.8000)  acc5: 66.4000 (66.4000)  time: 13.7012  data: 13.6766  max mem: 2905
Test:  [10/50]  eta: 0:01:33  loss: 2.6871 (2.6018)  acc1: 45.6000 (44.8000)  acc5: 70.4000 (69.5273)  time: 2.3336  data: 2.3127  max mem: 2905
Test:  [20/50]  eta: 0:00:55  loss: 2.8070 (2.8404)  acc1: 39.2000 (41.0286)  acc5: 66.4000 (66.0191)  time: 1.2635  data: 1.2439  max mem: 2905
Test:  [30/50]  eta: 0:00:33  loss: 2.9869 (2.8374)  acc1: 38.4000 (40.6194)  acc5: 64.0000 (65.8065)  time: 1.2906  data: 1.2722  max mem: 2905
Test:  [40/50]  eta: 0:00:13  loss: 2.9134 (2.8546)  acc1: 38.4000 (40.2732)  acc5: 64.0000 (65.3073)  time: 0.8778  data: 0.8594  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.8201 (2.8462)  acc1: 38.4000 (40.4320)  acc5: 64.8000 (65.5040)  time: 0.7974  data: 0.7773  max mem: 2905
Test: Total time: 0:00:57 (1.1475 s / it)
* Acc@1 40.948 Acc@5 66.198 loss 2.818
Accuracy of the model on the 50000 test images: 40.9%
Max accuracy: 50.54%
Epoch: [46]  [  0/625]  eta: 3:38:00  lr: 0.003916  min_lr: 0.003916  loss: 3.4364 (3.4364)  class_acc: 0.4492 (0.4492)  weight_decay: 0.0500 (0.0500)  time: 20.9293  data: 16.6041  max mem: 2905
Epoch: [46]  [200/625]  eta: 0:14:44  lr: 0.003913  min_lr: 0.003913  loss: 3.4116 (3.3715)  class_acc: 0.4375 (0.4477)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.8070 (2.6428)  time: 1.9882  data: 0.0006  max mem: 2905
Epoch: [46]  [400/625]  eta: 0:07:30  lr: 0.003911  min_lr: 0.003911  loss: 3.3764 (3.3810)  class_acc: 0.4375 (0.4464)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.5905 (2.3596)  time: 1.9217  data: 0.0006  max mem: 2905
Epoch: [46]  [600/625]  eta: 0:00:49  lr: 0.003909  min_lr: 0.003909  loss: 3.4171 (3.3907)  class_acc: 0.4258 (0.4433)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3015 (2.3302)  time: 1.9684  data: 0.0006  max mem: 2905
Epoch: [46]  [624/625]  eta: 0:00:01  lr: 0.003909  min_lr: 0.003909  loss: 3.4027 (3.3911)  class_acc: 0.4453 (0.4435)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7630 (2.3911)  time: 0.7131  data: 0.0013  max mem: 2905
Epoch: [46] Total time: 0:20:17 (1.9483 s / it)
Averaged stats: lr: 0.003909  min_lr: 0.003909  loss: 3.4027 (3.3825)  class_acc: 0.4453 (0.4445)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7630 (2.3911)
Test:  [ 0/50]  eta: 0:10:19  loss: 2.4703 (2.4703)  acc1: 48.0000 (48.0000)  acc5: 72.8000 (72.8000)  time: 12.3987  data: 12.3723  max mem: 2905
Test:  [10/50]  eta: 0:01:17  loss: 2.4856 (2.4961)  acc1: 46.4000 (46.9091)  acc5: 71.2000 (71.2000)  time: 1.9487  data: 1.9267  max mem: 2905
Test:  [20/50]  eta: 0:00:44  loss: 2.5880 (2.5977)  acc1: 44.0000 (44.8762)  acc5: 69.6000 (70.2476)  time: 0.9253  data: 0.9051  max mem: 2905
Test:  [30/50]  eta: 0:00:25  loss: 2.6890 (2.5957)  acc1: 42.4000 (44.6710)  acc5: 68.8000 (70.1677)  time: 0.8834  data: 0.8637  max mem: 2905
Test:  [40/50]  eta: 0:00:10  loss: 2.7281 (2.6489)  acc1: 39.2000 (43.4146)  acc5: 67.2000 (69.2293)  time: 0.6471  data: 0.6269  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 2.7281 (2.6545)  acc1: 40.8000 (43.5680)  acc5: 67.2000 (68.9600)  time: 0.6357  data: 0.6158  max mem: 2905
Test: Total time: 0:00:44 (0.8898 s / it)
* Acc@1 44.068 Acc@5 69.208 loss 2.618
Accuracy of the model on the 50000 test images: 44.1%
Max accuracy: 50.54%
Epoch: [47]  [  0/625]  eta: 3:52:41  lr: 0.003909  min_lr: 0.003909  loss: 3.1954 (3.1954)  class_acc: 0.4492 (0.4492)  weight_decay: 0.0500 (0.0500)  time: 22.3390  data: 18.0115  max mem: 2905
Epoch: [47]  [200/625]  eta: 0:14:28  lr: 0.003907  min_lr: 0.003907  loss: 3.3203 (3.3673)  class_acc: 0.4492 (0.4469)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1779 (2.3875)  time: 1.9263  data: 0.0006  max mem: 2905
Epoch: [47]  [400/625]  eta: 0:07:27  lr: 0.003905  min_lr: 0.003905  loss: 3.3434 (3.3741)  class_acc: 0.4414 (0.4463)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8667 (2.2675)  time: 1.9413  data: 0.0013  max mem: 2905
Epoch: [47]  [600/625]  eta: 0:00:50  lr: 0.003902  min_lr: 0.003902  loss: 3.3939 (3.3779)  class_acc: 0.4336 (0.4456)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8650 (inf)  time: 1.9792  data: 0.0008  max mem: 2905
Epoch: [47]  [624/625]  eta: 0:00:01  lr: 0.003902  min_lr: 0.003902  loss: 3.4238 (3.3784)  class_acc: 0.4492 (0.4457)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8650 (inf)  time: 0.8030  data: 0.0022  max mem: 2905
Epoch: [47] Total time: 0:20:26 (1.9628 s / it)
Averaged stats: lr: 0.003902  min_lr: 0.003902  loss: 3.4238 (3.3762)  class_acc: 0.4492 (0.4452)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8650 (inf)
Test:  [ 0/50]  eta: 0:11:01  loss: 2.3562 (2.3562)  acc1: 45.6000 (45.6000)  acc5: 74.4000 (74.4000)  time: 13.2286  data: 13.1987  max mem: 2905
Test:  [10/50]  eta: 0:01:31  loss: 2.4464 (2.3885)  acc1: 51.2000 (50.6909)  acc5: 71.2000 (72.5091)  time: 2.2831  data: 2.2626  max mem: 2905
Test:  [20/50]  eta: 0:00:54  loss: 2.5119 (2.5128)  acc1: 48.0000 (46.8952)  acc5: 70.4000 (71.2000)  time: 1.2430  data: 1.2241  max mem: 2905
Test:  [30/50]  eta: 0:00:32  loss: 2.6188 (2.5403)  acc1: 44.0000 (46.1419)  acc5: 70.4000 (70.7613)  time: 1.2836  data: 1.2650  max mem: 2905
Test:  [40/50]  eta: 0:00:13  loss: 2.5691 (2.5498)  acc1: 44.0000 (46.0878)  acc5: 70.4000 (70.3610)  time: 0.9235  data: 0.9041  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.5691 (2.5482)  acc1: 44.0000 (45.8240)  acc5: 71.2000 (70.6240)  time: 0.8419  data: 0.8226  max mem: 2905
Test: Total time: 0:00:57 (1.1471 s / it)
* Acc@1 46.410 Acc@5 71.536 loss 2.492
Accuracy of the model on the 50000 test images: 46.4%
Max accuracy: 50.54%
Epoch: [48]  [  0/625]  eta: 3:52:03  lr: 0.003902  min_lr: 0.003902  loss: 3.1786 (3.1786)  class_acc: 0.5156 (0.5156)  weight_decay: 0.0500 (0.0500)  time: 22.2780  data: 20.1751  max mem: 2905
Epoch: [48]  [200/625]  eta: 0:14:21  lr: 0.003900  min_lr: 0.003900  loss: 3.3386 (3.3467)  class_acc: 0.4414 (0.4495)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4963 (2.2951)  time: 1.7452  data: 0.3338  max mem: 2905
Epoch: [48]  [400/625]  eta: 0:07:31  lr: 0.003898  min_lr: 0.003898  loss: 3.3932 (3.3649)  class_acc: 0.4375 (0.4464)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.3706 (2.1958)  time: 1.9722  data: 0.0006  max mem: 2905
Epoch: [48]  [600/625]  eta: 0:00:49  lr: 0.003895  min_lr: 0.003895  loss: 3.3079 (3.3682)  class_acc: 0.4609 (0.4462)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2261 (2.2688)  time: 2.0331  data: 0.4061  max mem: 2905
Epoch: [48]  [624/625]  eta: 0:00:01  lr: 0.003895  min_lr: 0.003895  loss: 3.4283 (3.3705)  class_acc: 0.4258 (0.4457)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7065 (2.2906)  time: 0.8636  data: 0.2198  max mem: 2905
Epoch: [48] Total time: 0:20:20 (1.9523 s / it)
Averaged stats: lr: 0.003895  min_lr: 0.003895  loss: 3.4283 (3.3713)  class_acc: 0.4258 (0.4466)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7065 (2.2906)
Test:  [ 0/50]  eta: 0:09:59  loss: 2.7328 (2.7328)  acc1: 37.6000 (37.6000)  acc5: 66.4000 (66.4000)  time: 11.9817  data: 11.9570  max mem: 2905
Test:  [10/50]  eta: 0:01:22  loss: 2.5133 (2.4935)  acc1: 49.6000 (48.1455)  acc5: 72.0000 (72.0727)  time: 2.0612  data: 2.0427  max mem: 2905
Test:  [20/50]  eta: 0:00:48  loss: 2.5796 (2.5711)  acc1: 44.8000 (46.0190)  acc5: 70.4000 (70.6667)  time: 1.1021  data: 1.0837  max mem: 2905
Test:  [30/50]  eta: 0:00:27  loss: 2.5841 (2.5604)  acc1: 44.8000 (46.0645)  acc5: 69.6000 (70.5290)  time: 1.0160  data: 0.9962  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 2.5841 (2.6177)  acc1: 45.6000 (45.2293)  acc5: 68.0000 (69.2878)  time: 0.5999  data: 0.5794  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 2.5875 (2.6161)  acc1: 46.4000 (45.0400)  acc5: 68.0000 (69.6640)  time: 0.5985  data: 0.5793  max mem: 2905
Test: Total time: 0:00:46 (0.9272 s / it)
* Acc@1 45.200 Acc@5 70.318 loss 2.603
Accuracy of the model on the 50000 test images: 45.2%
Max accuracy: 50.54%
Epoch: [49]  [  0/625]  eta: 3:44:40  lr: 0.003895  min_lr: 0.003895  loss: 3.2399 (3.2399)  class_acc: 0.4609 (0.4609)  weight_decay: 0.0500 (0.0500)  time: 21.5688  data: 17.3616  max mem: 2905
Epoch: [49]  [200/625]  eta: 0:14:05  lr: 0.003893  min_lr: 0.003893  loss: 3.3509 (3.3598)  class_acc: 0.4492 (0.4500)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2987 (2.4050)  time: 1.7431  data: 0.0009  max mem: 2905
Epoch: [49]  [400/625]  eta: 0:07:10  lr: 0.003890  min_lr: 0.003890  loss: 3.3827 (3.3656)  class_acc: 0.4453 (0.4474)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.0852 (2.3308)  time: 1.7235  data: 0.0007  max mem: 2905
Epoch: [49]  [600/625]  eta: 0:00:47  lr: 0.003888  min_lr: 0.003888  loss: 3.3552 (3.3663)  class_acc: 0.4375 (0.4469)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3929 (2.3984)  time: 1.9828  data: 0.0009  max mem: 2905
Epoch: [49]  [624/625]  eta: 0:00:01  lr: 0.003888  min_lr: 0.003888  loss: 3.4198 (3.3683)  class_acc: 0.4297 (0.4465)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7160 (2.3695)  time: 0.7373  data: 0.0339  max mem: 2905
Epoch: [49] Total time: 0:19:39 (1.8867 s / it)
Averaged stats: lr: 0.003888  min_lr: 0.003888  loss: 3.4198 (3.3642)  class_acc: 0.4297 (0.4474)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7160 (2.3695)
Test:  [ 0/50]  eta: 0:10:45  loss: 2.0867 (2.0867)  acc1: 50.4000 (50.4000)  acc5: 79.2000 (79.2000)  time: 12.9013  data: 12.8659  max mem: 2905
Test:  [10/50]  eta: 0:01:30  loss: 2.2798 (2.2423)  acc1: 51.2000 (51.7818)  acc5: 74.4000 (74.6182)  time: 2.2561  data: 2.2347  max mem: 2905
Test:  [20/50]  eta: 0:00:53  loss: 2.4046 (2.4170)  acc1: 47.2000 (48.0381)  acc5: 72.0000 (72.9524)  time: 1.2368  data: 1.2163  max mem: 2905
Test:  [30/50]  eta: 0:00:30  loss: 2.5913 (2.4262)  acc1: 44.8000 (47.4839)  acc5: 71.2000 (73.0581)  time: 1.1101  data: 1.0898  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 2.4528 (2.4268)  acc1: 46.4000 (47.6488)  acc5: 71.2000 (72.7805)  time: 0.6835  data: 0.6651  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.3834 (2.4297)  acc1: 46.4000 (47.3440)  acc5: 73.6000 (72.9440)  time: 0.5962  data: 0.5773  max mem: 2905
Test: Total time: 0:00:52 (1.0450 s / it)
* Acc@1 48.114 Acc@5 73.030 loss 2.409
Accuracy of the model on the 50000 test images: 48.1%
Max accuracy: 50.54%
Epoch: [50]  [  0/625]  eta: 4:07:52  lr: 0.003888  min_lr: 0.003888  loss: 3.3523 (3.3523)  class_acc: 0.4023 (0.4023)  weight_decay: 0.0500 (0.0500)  time: 23.7968  data: 19.1362  max mem: 2905
Epoch: [50]  [200/625]  eta: 0:13:33  lr: 0.003885  min_lr: 0.003885  loss: 3.3637 (3.3469)  class_acc: 0.4492 (0.4533)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.0410 (2.2378)  time: 1.7296  data: 0.0336  max mem: 2905
Epoch: [50]  [400/625]  eta: 0:07:13  lr: 0.003883  min_lr: 0.003883  loss: 3.3803 (3.3584)  class_acc: 0.4336 (0.4501)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2490 (2.2780)  time: 2.0968  data: 0.1872  max mem: 2905
Epoch: [50]  [600/625]  eta: 0:00:47  lr: 0.003881  min_lr: 0.003881  loss: 3.4428 (3.3672)  class_acc: 0.4453 (0.4478)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6783 (2.2415)  time: 1.9913  data: 0.0180  max mem: 2905
Epoch: [50]  [624/625]  eta: 0:00:01  lr: 0.003880  min_lr: 0.003880  loss: 3.3217 (3.3662)  class_acc: 0.4375 (0.4478)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9183 (2.2416)  time: 0.7108  data: 0.0150  max mem: 2905
Epoch: [50] Total time: 0:19:27 (1.8674 s / it)
Averaged stats: lr: 0.003880  min_lr: 0.003880  loss: 3.3217 (3.3603)  class_acc: 0.4375 (0.4487)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9183 (2.2416)
Test:  [ 0/50]  eta: 0:09:06  loss: 2.3292 (2.3292)  acc1: 46.4000 (46.4000)  acc5: 75.2000 (75.2000)  time: 10.9310  data: 10.9059  max mem: 2905
Test:  [10/50]  eta: 0:01:12  loss: 2.3329 (2.4272)  acc1: 47.2000 (47.7091)  acc5: 75.2000 (72.3636)  time: 1.8244  data: 1.8051  max mem: 2905
Test:  [20/50]  eta: 0:00:42  loss: 2.6642 (2.5996)  acc1: 44.0000 (43.5429)  acc5: 68.0000 (69.4095)  time: 0.9403  data: 0.9215  max mem: 2905
Test:  [30/50]  eta: 0:00:25  loss: 2.6907 (2.5875)  acc1: 42.4000 (44.2065)  acc5: 68.0000 (69.6516)  time: 0.9788  data: 0.9597  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 2.5704 (2.6325)  acc1: 44.8000 (43.5512)  acc5: 68.0000 (68.8585)  time: 0.8873  data: 0.8681  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.7138 (2.6392)  acc1: 44.0000 (43.7440)  acc5: 67.2000 (68.7840)  time: 0.5399  data: 0.5206  max mem: 2905
Test: Total time: 0:00:50 (1.0120 s / it)
* Acc@1 44.782 Acc@5 69.550 loss 2.608
Accuracy of the model on the 50000 test images: 44.8%
Max accuracy: 50.54%
Epoch: [51]  [  0/625]  eta: 3:59:38  lr: 0.003880  min_lr: 0.003880  loss: 3.0909 (3.0909)  class_acc: 0.4844 (0.4844)  weight_decay: 0.0500 (0.0500)  time: 23.0050  data: 17.1572  max mem: 2905
Epoch: [51]  [200/625]  eta: 0:13:39  lr: 0.003878  min_lr: 0.003878  loss: 3.3334 (3.3462)  class_acc: 0.4492 (0.4511)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1487 (2.3534)  time: 1.7712  data: 0.2478  max mem: 2905
Epoch: [51]  [400/625]  eta: 0:07:15  lr: 0.003875  min_lr: 0.003875  loss: 3.3337 (3.3544)  class_acc: 0.4609 (0.4518)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7774 (2.3754)  time: 1.8824  data: 0.0014  max mem: 2905
Epoch: [51]  [600/625]  eta: 0:00:49  lr: 0.003873  min_lr: 0.003873  loss: 3.3345 (3.3542)  class_acc: 0.4492 (0.4509)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.4334 (2.3757)  time: 2.0715  data: 0.0067  max mem: 2905
Epoch: [51]  [624/625]  eta: 0:00:01  lr: 0.003873  min_lr: 0.003873  loss: 3.3649 (3.3550)  class_acc: 0.4531 (0.4509)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.4907 (2.3431)  time: 1.1988  data: 0.0014  max mem: 2905
Epoch: [51] Total time: 0:20:08 (1.9338 s / it)
Averaged stats: lr: 0.003873  min_lr: 0.003873  loss: 3.3649 (3.3532)  class_acc: 0.4531 (0.4504)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.4907 (2.3431)
Test:  [ 0/50]  eta: 0:09:50  loss: 2.2951 (2.2951)  acc1: 44.8000 (44.8000)  acc5: 77.6000 (77.6000)  time: 11.8174  data: 11.7762  max mem: 2905
Test:  [10/50]  eta: 0:01:19  loss: 2.2951 (2.3385)  acc1: 50.4000 (50.8364)  acc5: 75.2000 (74.1818)  time: 1.9801  data: 1.9595  max mem: 2905
Test:  [20/50]  eta: 0:00:47  loss: 2.4481 (2.4977)  acc1: 47.2000 (46.6667)  acc5: 71.2000 (72.2286)  time: 1.0553  data: 1.0363  max mem: 2905
Test:  [30/50]  eta: 0:00:28  loss: 2.6484 (2.4995)  acc1: 43.2000 (46.5032)  acc5: 69.6000 (71.8968)  time: 1.0884  data: 1.0690  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 2.6053 (2.5626)  acc1: 43.2000 (45.4244)  acc5: 68.0000 (70.8683)  time: 0.9251  data: 0.9045  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.5410 (2.5572)  acc1: 45.6000 (45.9040)  acc5: 70.4000 (70.9440)  time: 0.8118  data: 0.7913  max mem: 2905
Test: Total time: 0:00:56 (1.1241 s / it)
* Acc@1 46.402 Acc@5 71.460 loss 2.520
Accuracy of the model on the 50000 test images: 46.4%
Max accuracy: 50.54%
Epoch: [52]  [  0/625]  eta: 3:43:35  lr: 0.003873  min_lr: 0.003873  loss: 3.3729 (3.3729)  class_acc: 0.4766 (0.4766)  weight_decay: 0.0500 (0.0500)  time: 21.4641  data: 21.3380  max mem: 2905
Epoch: [52]  [200/625]  eta: 0:14:33  lr: 0.003870  min_lr: 0.003870  loss: 3.3844 (3.3343)  class_acc: 0.4375 (0.4534)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8153 (2.3459)  time: 1.9381  data: 1.3273  max mem: 2905
Epoch: [52]  [400/625]  eta: 0:07:29  lr: 0.003867  min_lr: 0.003867  loss: 3.3677 (3.3465)  class_acc: 0.4375 (0.4509)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.4532 (2.4347)  time: 1.9345  data: 0.0007  max mem: 2905
Epoch: [52]  [600/625]  eta: 0:00:50  lr: 0.003865  min_lr: 0.003865  loss: 3.3735 (3.3534)  class_acc: 0.4336 (0.4499)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8741 (2.3795)  time: 2.0584  data: 0.0008  max mem: 2905
Epoch: [52]  [624/625]  eta: 0:00:01  lr: 0.003865  min_lr: 0.003865  loss: 3.3151 (3.3529)  class_acc: 0.4531 (0.4498)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3008 (2.4096)  time: 0.7025  data: 0.0040  max mem: 2905
Epoch: [52] Total time: 0:20:21 (1.9546 s / it)
Averaged stats: lr: 0.003865  min_lr: 0.003865  loss: 3.3151 (3.3513)  class_acc: 0.4531 (0.4506)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3008 (2.4096)
Test:  [ 0/50]  eta: 0:10:44  loss: 2.3972 (2.3972)  acc1: 46.4000 (46.4000)  acc5: 76.0000 (76.0000)  time: 12.8923  data: 12.8580  max mem: 2905
Test:  [10/50]  eta: 0:01:29  loss: 2.3972 (2.4438)  acc1: 49.6000 (49.1636)  acc5: 72.8000 (72.4364)  time: 2.2326  data: 2.2132  max mem: 2905
Test:  [20/50]  eta: 0:00:54  loss: 2.5716 (2.6270)  acc1: 46.4000 (44.4952)  acc5: 68.8000 (69.5619)  time: 1.2482  data: 1.2296  max mem: 2905
Test:  [30/50]  eta: 0:00:30  loss: 2.7839 (2.6609)  acc1: 38.4000 (43.3290)  acc5: 68.0000 (69.2129)  time: 1.1617  data: 1.1430  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 2.7189 (2.6865)  acc1: 40.0000 (42.8488)  acc5: 68.0000 (68.7610)  time: 0.6817  data: 0.6640  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.6566 (2.6838)  acc1: 40.0000 (42.8000)  acc5: 67.2000 (68.6080)  time: 0.5759  data: 0.5584  max mem: 2905
Test: Total time: 0:00:52 (1.0487 s / it)
* Acc@1 43.642 Acc@5 68.832 loss 2.653
Accuracy of the model on the 50000 test images: 43.6%
Max accuracy: 50.54%
Epoch: [53]  [  0/625]  eta: 3:38:28  lr: 0.003865  min_lr: 0.003865  loss: 3.3371 (3.3371)  class_acc: 0.4688 (0.4688)  weight_decay: 0.0500 (0.0500)  time: 20.9733  data: 20.7793  max mem: 2905
Epoch: [53]  [200/625]  eta: 0:14:13  lr: 0.003862  min_lr: 0.003862  loss: 3.2799 (3.3253)  class_acc: 0.4492 (0.4560)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7506 (2.3380)  time: 1.9518  data: 1.3974  max mem: 2905
Epoch: [53]  [400/625]  eta: 0:07:25  lr: 0.003859  min_lr: 0.003859  loss: 3.3510 (3.3319)  class_acc: 0.4453 (0.4556)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6383 (2.4065)  time: 1.9923  data: 1.7855  max mem: 2905
Epoch: [53]  [600/625]  eta: 0:00:48  lr: 0.003857  min_lr: 0.003857  loss: 3.3991 (3.3381)  class_acc: 0.4180 (0.4539)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0424 (2.3773)  time: 1.7336  data: 0.9385  max mem: 2905
Epoch: [53]  [624/625]  eta: 0:00:01  lr: 0.003856  min_lr: 0.003856  loss: 3.3224 (3.3387)  class_acc: 0.4414 (0.4538)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8395 (2.3796)  time: 0.6873  data: 0.1658  max mem: 2905
Epoch: [53] Total time: 0:19:50 (1.9047 s / it)
Averaged stats: lr: 0.003856  min_lr: 0.003856  loss: 3.3224 (3.3406)  class_acc: 0.4414 (0.4530)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8395 (2.3796)
Test:  [ 0/50]  eta: 0:09:10  loss: 2.1193 (2.1193)  acc1: 54.4000 (54.4000)  acc5: 77.6000 (77.6000)  time: 11.0062  data: 10.9774  max mem: 2905
Test:  [10/50]  eta: 0:01:19  loss: 2.1193 (2.1892)  acc1: 52.0000 (51.8545)  acc5: 76.8000 (76.0000)  time: 1.9799  data: 1.9603  max mem: 2905
Test:  [20/50]  eta: 0:00:46  loss: 2.3489 (2.3282)  acc1: 48.0000 (49.1810)  acc5: 75.2000 (74.1714)  time: 1.0845  data: 1.0654  max mem: 2905
Test:  [30/50]  eta: 0:00:26  loss: 2.3489 (2.3033)  acc1: 48.8000 (50.1677)  acc5: 73.6000 (74.6581)  time: 0.9899  data: 0.9690  max mem: 2905
Test:  [40/50]  eta: 0:00:10  loss: 2.3220 (2.3416)  acc1: 48.8000 (48.9366)  acc5: 74.4000 (73.8732)  time: 0.5976  data: 0.5760  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 2.3411 (2.3414)  acc1: 47.2000 (49.0880)  acc5: 73.6000 (73.8720)  time: 0.5045  data: 0.4844  max mem: 2905
Test: Total time: 0:00:45 (0.9145 s / it)
* Acc@1 50.186 Acc@5 74.632 loss 2.302
Accuracy of the model on the 50000 test images: 50.2%
Max accuracy: 50.54%
Epoch: [54]  [  0/625]  eta: 3:28:15  lr: 0.003856  min_lr: 0.003856  loss: 3.5111 (3.5111)  class_acc: 0.4219 (0.4219)  weight_decay: 0.0500 (0.0500)  time: 19.9926  data: 19.8575  max mem: 2905
Epoch: [54]  [200/625]  eta: 0:13:54  lr: 0.003854  min_lr: 0.003854  loss: 3.3217 (3.3241)  class_acc: 0.4648 (0.4572)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9949 (2.3172)  time: 1.9306  data: 1.7631  max mem: 2905
Epoch: [54]  [400/625]  eta: 0:07:08  lr: 0.003851  min_lr: 0.003851  loss: 3.3417 (3.3287)  class_acc: 0.4492 (0.4559)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.9670 (inf)  time: 1.7860  data: 1.6256  max mem: 2905
Epoch: [54]  [600/625]  eta: 0:00:47  lr: 0.003848  min_lr: 0.003848  loss: 3.3792 (3.3330)  class_acc: 0.4414 (0.4550)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0976 (inf)  time: 1.7896  data: 0.5088  max mem: 2905
Epoch: [54]  [624/625]  eta: 0:00:01  lr: 0.003848  min_lr: 0.003848  loss: 3.3487 (3.3338)  class_acc: 0.4453 (0.4551)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2949 (inf)  time: 0.8153  data: 0.3852  max mem: 2905
Epoch: [54] Total time: 0:19:25 (1.8656 s / it)
Averaged stats: lr: 0.003848  min_lr: 0.003848  loss: 3.3487 (3.3365)  class_acc: 0.4453 (0.4541)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2949 (inf)
Test:  [ 0/50]  eta: 0:10:09  loss: 2.6353 (2.6353)  acc1: 44.0000 (44.0000)  acc5: 68.0000 (68.0000)  time: 12.1905  data: 12.1571  max mem: 2905
Test:  [10/50]  eta: 0:01:15  loss: 2.6075 (2.5707)  acc1: 46.4000 (47.4182)  acc5: 68.0000 (69.5273)  time: 1.8995  data: 1.8780  max mem: 2905
Test:  [20/50]  eta: 0:00:43  loss: 2.6569 (2.6665)  acc1: 44.8000 (44.3810)  acc5: 68.0000 (68.8381)  time: 0.9079  data: 0.8879  max mem: 2905
Test:  [30/50]  eta: 0:00:26  loss: 2.7139 (2.6335)  acc1: 41.6000 (44.5936)  acc5: 68.8000 (69.1097)  time: 0.9731  data: 0.9534  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 2.5953 (2.6445)  acc1: 41.6000 (44.6829)  acc5: 68.8000 (69.1122)  time: 0.9164  data: 0.8971  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.6022 (2.6296)  acc1: 41.6000 (44.9280)  acc5: 68.8000 (69.2960)  time: 0.6418  data: 0.6237  max mem: 2905
Test: Total time: 0:00:53 (1.0665 s / it)
* Acc@1 45.138 Acc@5 69.818 loss 2.596
Accuracy of the model on the 50000 test images: 45.1%
Max accuracy: 50.54%
Epoch: [55]  [  0/625]  eta: 3:45:14  lr: 0.003848  min_lr: 0.003848  loss: 3.3281 (3.3281)  class_acc: 0.4336 (0.4336)  weight_decay: 0.0500 (0.0500)  time: 21.6229  data: 18.5020  max mem: 2905
Epoch: [55]  [200/625]  eta: 0:14:17  lr: 0.003845  min_lr: 0.003845  loss: 3.2878 (3.3199)  class_acc: 0.4609 (0.4595)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8169 (2.3169)  time: 1.9516  data: 0.0713  max mem: 2905
Epoch: [55]  [400/625]  eta: 0:07:21  lr: 0.003842  min_lr: 0.003842  loss: 3.3595 (3.3283)  class_acc: 0.4531 (0.4565)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1946 (2.3024)  time: 1.8699  data: 0.3091  max mem: 2905
Epoch: [55]  [600/625]  eta: 0:00:48  lr: 0.003839  min_lr: 0.003839  loss: 3.2922 (3.3343)  class_acc: 0.4570 (0.4550)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0148 (2.3592)  time: 2.0353  data: 0.2675  max mem: 2905
Epoch: [55]  [624/625]  eta: 0:00:01  lr: 0.003839  min_lr: 0.003839  loss: 3.3192 (3.3334)  class_acc: 0.4609 (0.4552)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9418 (2.3589)  time: 0.7735  data: 0.0180  max mem: 2905
Epoch: [55] Total time: 0:19:56 (1.9138 s / it)
Averaged stats: lr: 0.003839  min_lr: 0.003839  loss: 3.3192 (3.3316)  class_acc: 0.4609 (0.4546)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9418 (2.3589)
Test:  [ 0/50]  eta: 0:10:13  loss: 2.5900 (2.5900)  acc1: 42.4000 (42.4000)  acc5: 72.0000 (72.0000)  time: 12.2720  data: 12.2428  max mem: 2905
Test:  [10/50]  eta: 0:01:22  loss: 2.4106 (2.4076)  acc1: 48.0000 (48.1455)  acc5: 72.0000 (72.5091)  time: 2.0528  data: 2.0318  max mem: 2905
Test:  [20/50]  eta: 0:00:43  loss: 2.4658 (2.5802)  acc1: 44.8000 (44.8762)  acc5: 68.8000 (69.7905)  time: 0.9145  data: 0.8941  max mem: 2905
Test:  [30/50]  eta: 0:00:24  loss: 2.6815 (2.5835)  acc1: 41.6000 (44.4645)  acc5: 68.0000 (69.4452)  time: 0.7759  data: 0.7548  max mem: 2905
Test:  [40/50]  eta: 0:00:10  loss: 2.5540 (2.6018)  acc1: 40.8000 (43.9415)  acc5: 70.4000 (69.4244)  time: 0.6095  data: 0.5892  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 2.5929 (2.6230)  acc1: 43.2000 (43.6800)  acc5: 69.6000 (69.1360)  time: 0.3771  data: 0.3575  max mem: 2905
Test: Total time: 0:00:45 (0.9085 s / it)
* Acc@1 44.820 Acc@5 69.716 loss 2.599
Accuracy of the model on the 50000 test images: 44.8%
Max accuracy: 50.54%
Epoch: [56]  [  0/625]  eta: 3:37:05  lr: 0.003839  min_lr: 0.003839  loss: 3.2067 (3.2067)  class_acc: 0.4727 (0.4727)  weight_decay: 0.0500 (0.0500)  time: 20.8406  data: 19.3918  max mem: 2905
Epoch: [56]  [200/625]  eta: 0:14:23  lr: 0.003836  min_lr: 0.003836  loss: 3.2960 (3.3139)  class_acc: 0.4609 (0.4597)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9604 (2.2977)  time: 1.9227  data: 0.1201  max mem: 2905
Epoch: [56]  [400/625]  eta: 0:07:30  lr: 0.003833  min_lr: 0.003833  loss: 3.3843 (3.3160)  class_acc: 0.4414 (0.4577)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.6814 (2.3021)  time: 1.9409  data: 0.0009  max mem: 2905
Epoch: [56]  [600/625]  eta: 0:00:49  lr: 0.003831  min_lr: 0.003831  loss: 3.3578 (3.3277)  class_acc: 0.4492 (0.4549)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7998 (2.4202)  time: 2.0004  data: 0.0007  max mem: 2905
Epoch: [56]  [624/625]  eta: 0:00:01  lr: 0.003830  min_lr: 0.003830  loss: 3.3479 (3.3292)  class_acc: 0.4375 (0.4547)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.5540 (2.4038)  time: 0.8399  data: 0.0325  max mem: 2905
Epoch: [56] Total time: 0:20:19 (1.9514 s / it)
Averaged stats: lr: 0.003830  min_lr: 0.003830  loss: 3.3479 (3.3284)  class_acc: 0.4375 (0.4555)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.5540 (2.4038)
Test:  [ 0/50]  eta: 0:11:16  loss: 3.0927 (3.0927)  acc1: 38.4000 (38.4000)  acc5: 63.2000 (63.2000)  time: 13.5395  data: 13.5018  max mem: 2905
Test:  [10/50]  eta: 0:01:25  loss: 2.7362 (2.7250)  acc1: 44.8000 (44.3636)  acc5: 68.0000 (67.6364)  time: 2.1336  data: 2.1126  max mem: 2905
Test:  [20/50]  eta: 0:00:45  loss: 2.9154 (2.8930)  acc1: 39.2000 (40.5714)  acc5: 65.6000 (65.1429)  time: 0.9039  data: 0.8844  max mem: 2905
Test:  [30/50]  eta: 0:00:25  loss: 2.9884 (2.9015)  acc1: 37.6000 (40.6452)  acc5: 63.2000 (64.8516)  time: 0.8217  data: 0.8028  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 2.8629 (2.9498)  acc1: 40.0000 (39.7073)  acc5: 58.4000 (63.6293)  time: 0.8414  data: 0.8234  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.9843 (2.9531)  acc1: 41.6000 (39.8240)  acc5: 61.6000 (63.8400)  time: 0.5726  data: 0.5548  max mem: 2905
Test: Total time: 0:00:51 (1.0302 s / it)
* Acc@1 40.296 Acc@5 64.894 loss 2.903
Accuracy of the model on the 50000 test images: 40.3%
Max accuracy: 50.54%
Epoch: [57]  [  0/625]  eta: 3:30:03  lr: 0.003830  min_lr: 0.003830  loss: 3.2730 (3.2730)  class_acc: 0.4570 (0.4570)  weight_decay: 0.0500 (0.0500)  time: 20.1652  data: 19.7166  max mem: 2905
Epoch: [57]  [200/625]  eta: 0:14:21  lr: 0.003827  min_lr: 0.003827  loss: 3.2875 (3.3136)  class_acc: 0.4609 (0.4594)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2192 (2.4328)  time: 1.9736  data: 0.0192  max mem: 2905
Epoch: [57]  [400/625]  eta: 0:07:28  lr: 0.003824  min_lr: 0.003824  loss: 3.2251 (3.3176)  class_acc: 0.4688 (0.4580)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.3674 (2.3841)  time: 1.9043  data: 0.0007  max mem: 2905
Epoch: [57]  [600/625]  eta: 0:00:49  lr: 0.003821  min_lr: 0.003821  loss: 3.3144 (3.3283)  class_acc: 0.4570 (0.4563)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2357 (2.3572)  time: 2.0957  data: 0.0006  max mem: 2905
Epoch: [57]  [624/625]  eta: 0:00:01  lr: 0.003821  min_lr: 0.003821  loss: 3.3472 (3.3292)  class_acc: 0.4492 (0.4561)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.4799 (2.3247)  time: 0.6861  data: 0.0013  max mem: 2905
Epoch: [57] Total time: 0:20:15 (1.9448 s / it)
Averaged stats: lr: 0.003821  min_lr: 0.003821  loss: 3.3472 (3.3227)  class_acc: 0.4492 (0.4567)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.4799 (2.3247)
Test:  [ 0/50]  eta: 0:10:09  loss: 2.5353 (2.5353)  acc1: 41.6000 (41.6000)  acc5: 73.6000 (73.6000)  time: 12.1871  data: 12.1532  max mem: 2905
Test:  [10/50]  eta: 0:01:25  loss: 2.5353 (2.5381)  acc1: 46.4000 (46.2545)  acc5: 73.6000 (72.0000)  time: 2.1339  data: 2.1129  max mem: 2905
Test:  [20/50]  eta: 0:00:50  loss: 2.6292 (2.6446)  acc1: 44.0000 (43.7333)  acc5: 70.4000 (70.6286)  time: 1.1516  data: 1.1311  max mem: 2905
Test:  [30/50]  eta: 0:00:30  loss: 2.6422 (2.6565)  acc1: 42.4000 (43.6903)  acc5: 68.8000 (69.9097)  time: 1.1590  data: 1.1388  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 2.7315 (2.6883)  acc1: 40.8000 (43.2000)  acc5: 66.4000 (69.1707)  time: 0.7926  data: 0.7714  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.7178 (2.6932)  acc1: 42.4000 (43.3920)  acc5: 66.4000 (69.0400)  time: 0.7077  data: 0.6872  max mem: 2905
Test: Total time: 0:00:51 (1.0361 s / it)
* Acc@1 43.756 Acc@5 68.748 loss 2.676
Accuracy of the model on the 50000 test images: 43.8%
Max accuracy: 50.54%
Epoch: [58]  [  0/625]  eta: 4:51:19  lr: 0.003821  min_lr: 0.003821  loss: 3.0344 (3.0344)  class_acc: 0.5273 (0.5273)  weight_decay: 0.0500 (0.0500)  time: 27.9675  data: 18.8793  max mem: 2905
Epoch: [58]  [200/625]  eta: 0:14:29  lr: 0.003818  min_lr: 0.003818  loss: 3.3546 (3.3013)  class_acc: 0.4570 (0.4609)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1308 (2.3499)  time: 2.1118  data: 0.0007  max mem: 2905
Epoch: [58]  [400/625]  eta: 0:07:27  lr: 0.003815  min_lr: 0.003815  loss: 3.2634 (3.3127)  class_acc: 0.4648 (0.4570)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.8615 (2.3693)  time: 1.8460  data: 0.0134  max mem: 2905
Epoch: [58]  [600/625]  eta: 0:00:49  lr: 0.003812  min_lr: 0.003812  loss: 3.3320 (3.3137)  class_acc: 0.4570 (0.4573)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9713 (2.3525)  time: 1.9947  data: 0.0399  max mem: 2905
Epoch: [58]  [624/625]  eta: 0:00:01  lr: 0.003812  min_lr: 0.003812  loss: 3.2654 (3.3127)  class_acc: 0.4609 (0.4576)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9461 (2.3378)  time: 0.5971  data: 0.0015  max mem: 2905
Epoch: [58] Total time: 0:20:23 (1.9581 s / it)
Averaged stats: lr: 0.003812  min_lr: 0.003812  loss: 3.2654 (3.3198)  class_acc: 0.4609 (0.4577)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9461 (2.3378)
Test:  [ 0/50]  eta: 0:10:21  loss: 2.7599 (2.7599)  acc1: 36.0000 (36.0000)  acc5: 65.6000 (65.6000)  time: 12.4284  data: 12.3951  max mem: 2905
Test:  [10/50]  eta: 0:01:21  loss: 2.4294 (2.4196)  acc1: 46.4000 (48.5091)  acc5: 72.8000 (71.6364)  time: 2.0481  data: 2.0272  max mem: 2905
Test:  [20/50]  eta: 0:00:47  loss: 2.5258 (2.6131)  acc1: 44.8000 (44.4190)  acc5: 67.2000 (69.0286)  time: 1.0430  data: 1.0240  max mem: 2905
Test:  [30/50]  eta: 0:00:28  loss: 2.7874 (2.6474)  acc1: 41.6000 (43.7936)  acc5: 65.6000 (69.0065)  time: 1.0583  data: 1.0383  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 2.7533 (2.7165)  acc1: 41.6000 (43.2000)  acc5: 66.4000 (67.9610)  time: 0.8515  data: 0.8311  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.7404 (2.7161)  acc1: 40.0000 (43.0720)  acc5: 66.4000 (68.1280)  time: 0.7604  data: 0.7415  max mem: 2905
Test: Total time: 0:00:54 (1.0802 s / it)
* Acc@1 43.572 Acc@5 68.920 loss 2.678
Accuracy of the model on the 50000 test images: 43.6%
Max accuracy: 50.54%
Epoch: [59]  [  0/625]  eta: 3:17:50  lr: 0.003812  min_lr: 0.003812  loss: 3.2325 (3.2325)  class_acc: 0.4688 (0.4688)  weight_decay: 0.0500 (0.0500)  time: 18.9934  data: 18.4385  max mem: 2905
Epoch: [59]  [200/625]  eta: 0:14:26  lr: 0.003809  min_lr: 0.003809  loss: 3.3065 (3.2989)  class_acc: 0.4570 (0.4624)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.4441 (2.4589)  time: 1.8814  data: 0.0467  max mem: 2905
Epoch: [59]  [400/625]  eta: 0:07:24  lr: 0.003805  min_lr: 0.003805  loss: 3.2682 (3.3057)  class_acc: 0.4492 (0.4601)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0477 (2.4738)  time: 1.9482  data: 0.0946  max mem: 2905
Epoch: [59]  [600/625]  eta: 0:00:48  lr: 0.003802  min_lr: 0.003802  loss: 3.3696 (3.3190)  class_acc: 0.4414 (0.4570)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6339 (2.4898)  time: 1.6085  data: 0.6684  max mem: 2905
Epoch: [59]  [624/625]  eta: 0:00:01  lr: 0.003802  min_lr: 0.003802  loss: 3.3636 (3.3204)  class_acc: 0.4570 (0.4570)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8166 (2.4651)  time: 0.8589  data: 0.3184  max mem: 2905
Epoch: [59] Total time: 0:20:02 (1.9239 s / it)
Averaged stats: lr: 0.003802  min_lr: 0.003802  loss: 3.3636 (3.3148)  class_acc: 0.4570 (0.4582)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8166 (2.4651)
Test:  [ 0/50]  eta: 0:09:50  loss: 2.5679 (2.5679)  acc1: 41.6000 (41.6000)  acc5: 68.0000 (68.0000)  time: 11.8145  data: 11.7875  max mem: 2905
Test:  [10/50]  eta: 0:01:19  loss: 2.5679 (2.6337)  acc1: 43.2000 (44.9455)  acc5: 68.8000 (69.0909)  time: 1.9793  data: 1.9586  max mem: 2905
Test:  [20/50]  eta: 0:00:46  loss: 2.8220 (2.7727)  acc1: 41.6000 (41.4857)  acc5: 67.2000 (66.9714)  time: 1.0306  data: 1.0111  max mem: 2905
Test:  [30/50]  eta: 0:00:26  loss: 2.9081 (2.8047)  acc1: 36.8000 (40.7226)  acc5: 65.6000 (66.1419)  time: 0.9852  data: 0.9664  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 2.8539 (2.8500)  acc1: 38.4000 (40.1951)  acc5: 64.0000 (65.8341)  time: 0.6664  data: 0.6475  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 2.8539 (2.8624)  acc1: 38.4000 (40.2560)  acc5: 63.2000 (65.6960)  time: 0.5486  data: 0.5293  max mem: 2905
Test: Total time: 0:00:48 (0.9692 s / it)
* Acc@1 41.166 Acc@5 66.114 loss 2.834
Accuracy of the model on the 50000 test images: 41.2%
Max accuracy: 50.54%
Epoch: [60]  [  0/625]  eta: 3:44:21  lr: 0.003802  min_lr: 0.003802  loss: 3.2687 (3.2687)  class_acc: 0.4648 (0.4648)  weight_decay: 0.0500 (0.0500)  time: 21.5383  data: 19.5985  max mem: 2905
Epoch: [60]  [200/625]  eta: 0:14:03  lr: 0.003799  min_lr: 0.003799  loss: 3.3259 (3.3044)  class_acc: 0.4492 (0.4620)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7361 (2.3662)  time: 2.1894  data: 0.1077  max mem: 2905
Epoch: [60]  [400/625]  eta: 0:07:19  lr: 0.003796  min_lr: 0.003796  loss: 3.3135 (3.3169)  class_acc: 0.4531 (0.4596)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7666 (2.3642)  time: 1.9194  data: 0.0060  max mem: 2905
Epoch: [60]  [600/625]  eta: 0:00:48  lr: 0.003793  min_lr: 0.003793  loss: 3.3658 (3.3245)  class_acc: 0.4531 (0.4576)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.8036 (2.4371)  time: 1.9693  data: 0.0098  max mem: 2905
Epoch: [60]  [624/625]  eta: 0:00:01  lr: 0.003792  min_lr: 0.003792  loss: 3.2929 (3.3241)  class_acc: 0.4531 (0.4576)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6227 (2.4382)  time: 0.4429  data: 0.0021  max mem: 2905
Epoch: [60] Total time: 0:19:49 (1.9039 s / it)
Averaged stats: lr: 0.003792  min_lr: 0.003792  loss: 3.2929 (3.3129)  class_acc: 0.4531 (0.4587)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6227 (2.4382)
Test:  [ 0/50]  eta: 0:10:19  loss: 2.4493 (2.4493)  acc1: 49.6000 (49.6000)  acc5: 72.8000 (72.8000)  time: 12.3956  data: 12.3685  max mem: 2905
Test:  [10/50]  eta: 0:01:24  loss: 2.4073 (2.2970)  acc1: 49.6000 (50.4000)  acc5: 72.8000 (74.6182)  time: 2.1071  data: 2.0854  max mem: 2905
Test:  [20/50]  eta: 0:00:49  loss: 2.4407 (2.4134)  acc1: 47.2000 (48.0000)  acc5: 71.2000 (73.1048)  time: 1.1267  data: 1.1064  max mem: 2905
Test:  [30/50]  eta: 0:00:29  loss: 2.4924 (2.4219)  acc1: 44.8000 (47.4323)  acc5: 71.2000 (72.6710)  time: 1.1551  data: 1.1348  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 2.5364 (2.4671)  acc1: 44.0000 (46.8878)  acc5: 71.2000 (71.5707)  time: 0.8972  data: 0.8773  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.6074 (2.4902)  acc1: 44.8000 (46.7360)  acc5: 69.6000 (71.6160)  time: 0.8044  data: 0.7864  max mem: 2905
Test: Total time: 0:00:53 (1.0737 s / it)
* Acc@1 46.664 Acc@5 71.960 loss 2.469
Accuracy of the model on the 50000 test images: 46.7%
Max accuracy: 50.54%
Epoch: [61]  [  0/625]  eta: 3:28:39  lr: 0.003792  min_lr: 0.003792  loss: 3.4290 (3.4290)  class_acc: 0.4180 (0.4180)  weight_decay: 0.0500 (0.0500)  time: 20.0314  data: 16.9222  max mem: 2905
Epoch: [61]  [200/625]  eta: 0:14:03  lr: 0.003789  min_lr: 0.003789  loss: 3.3371 (3.2947)  class_acc: 0.4609 (0.4648)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.3305 (inf)  time: 1.7724  data: 0.0008  max mem: 2905
Epoch: [61]  [400/625]  eta: 0:07:18  lr: 0.003786  min_lr: 0.003786  loss: 3.3204 (3.3116)  class_acc: 0.4570 (0.4608)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.5571 (inf)  time: 1.9372  data: 0.0364  max mem: 2905
Epoch: [61]  [600/625]  eta: 0:00:48  lr: 0.003782  min_lr: 0.003782  loss: 3.3354 (3.3174)  class_acc: 0.4570 (0.4592)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0327 (inf)  time: 1.9096  data: 0.0569  max mem: 2905
Epoch: [61]  [624/625]  eta: 0:00:01  lr: 0.003782  min_lr: 0.003782  loss: 3.3305 (3.3174)  class_acc: 0.4453 (0.4591)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0398 (inf)  time: 0.9253  data: 0.1530  max mem: 2905
Epoch: [61] Total time: 0:19:49 (1.9027 s / it)
Averaged stats: lr: 0.003782  min_lr: 0.003782  loss: 3.3305 (3.3062)  class_acc: 0.4453 (0.4601)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0398 (inf)
Test:  [ 0/50]  eta: 0:10:14  loss: 2.6603 (2.6603)  acc1: 42.4000 (42.4000)  acc5: 74.4000 (74.4000)  time: 12.2876  data: 12.2547  max mem: 2905
Test:  [10/50]  eta: 0:01:24  loss: 2.4463 (2.4569)  acc1: 46.4000 (47.4182)  acc5: 72.8000 (73.2364)  time: 2.1033  data: 2.0830  max mem: 2905
Test:  [20/50]  eta: 0:00:49  loss: 2.5358 (2.6238)  acc1: 45.6000 (44.0762)  acc5: 69.6000 (70.4381)  time: 1.1224  data: 1.1031  max mem: 2905
Test:  [30/50]  eta: 0:00:29  loss: 2.7010 (2.6093)  acc1: 41.6000 (44.0516)  acc5: 68.0000 (70.0129)  time: 1.1473  data: 1.1284  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 2.5750 (2.6176)  acc1: 44.0000 (44.1951)  acc5: 68.8000 (69.8341)  time: 0.8082  data: 0.7889  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.6903 (2.6433)  acc1: 43.2000 (43.9840)  acc5: 67.2000 (69.2800)  time: 0.7064  data: 0.6862  max mem: 2905
Test: Total time: 0:00:51 (1.0300 s / it)
* Acc@1 44.604 Acc@5 69.548 loss 2.620
Accuracy of the model on the 50000 test images: 44.6%
Max accuracy: 50.54%
Epoch: [62]  [  0/625]  eta: 3:34:36  lr: 0.003782  min_lr: 0.003782  loss: 3.2953 (3.2953)  class_acc: 0.4727 (0.4727)  weight_decay: 0.0500 (0.0500)  time: 20.6029  data: 18.2351  max mem: 2905
Epoch: [62]  [200/625]  eta: 0:14:20  lr: 0.003779  min_lr: 0.003779  loss: 3.3156 (3.2916)  class_acc: 0.4492 (0.4614)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9701 (2.4303)  time: 2.0279  data: 0.0635  max mem: 2905
Epoch: [62]  [400/625]  eta: 0:07:30  lr: 0.003775  min_lr: 0.003775  loss: 3.2369 (3.2940)  class_acc: 0.4766 (0.4630)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.5507 (2.3717)  time: 2.1138  data: 0.0062  max mem: 2905
Epoch: [62]  [600/625]  eta: 0:00:50  lr: 0.003772  min_lr: 0.003772  loss: 3.3230 (3.3032)  class_acc: 0.4609 (0.4611)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9138 (2.3462)  time: 1.9723  data: 0.0007  max mem: 2905
Epoch: [62]  [624/625]  eta: 0:00:01  lr: 0.003772  min_lr: 0.003772  loss: 3.3741 (3.3057)  class_acc: 0.4492 (0.4608)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.4766 (2.4026)  time: 0.7307  data: 0.0019  max mem: 2905
Epoch: [62] Total time: 0:20:24 (1.9587 s / it)
Averaged stats: lr: 0.003772  min_lr: 0.003772  loss: 3.3741 (3.3025)  class_acc: 0.4492 (0.4614)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.4766 (2.4026)
Test:  [ 0/50]  eta: 0:09:41  loss: 4.1889 (4.1889)  acc1: 24.8000 (24.8000)  acc5: 44.8000 (44.8000)  time: 11.6306  data: 11.5982  max mem: 2905
Test:  [10/50]  eta: 0:01:23  loss: 4.2697 (4.3001)  acc1: 23.2000 (23.2000)  acc5: 43.2000 (41.8909)  time: 2.0929  data: 2.0735  max mem: 2905
Test:  [20/50]  eta: 0:00:51  loss: 4.2838 (4.2976)  acc1: 21.6000 (22.4762)  acc5: 43.2000 (43.3143)  time: 1.2279  data: 1.2095  max mem: 2905
Test:  [30/50]  eta: 0:00:31  loss: 4.2838 (4.2823)  acc1: 21.6000 (23.1226)  acc5: 44.8000 (43.6645)  time: 1.2612  data: 1.2428  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 4.3064 (4.3028)  acc1: 22.4000 (22.9463)  acc5: 44.8000 (43.3171)  time: 0.8346  data: 0.8143  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 4.1999 (4.2837)  acc1: 23.2000 (23.5200)  acc5: 42.4000 (43.5360)  time: 0.6891  data: 0.6682  max mem: 2905
Test: Total time: 0:00:54 (1.0802 s / it)
* Acc@1 23.526 Acc@5 44.086 loss 4.261
Accuracy of the model on the 50000 test images: 23.5%
Max accuracy: 50.54%
Epoch: [63]  [  0/625]  eta: 3:55:24  lr: 0.003772  min_lr: 0.003772  loss: 3.0685 (3.0685)  class_acc: 0.4961 (0.4961)  weight_decay: 0.0500 (0.0500)  time: 22.5997  data: 17.5528  max mem: 2905
Epoch: [63]  [200/625]  eta: 0:15:04  lr: 0.003768  min_lr: 0.003768  loss: 3.2956 (3.2822)  class_acc: 0.4688 (0.4638)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.6090 (2.1121)  time: 2.0489  data: 0.0008  max mem: 2905
Epoch: [63]  [400/625]  eta: 0:07:41  lr: 0.003765  min_lr: 0.003765  loss: 3.3249 (3.2900)  class_acc: 0.4375 (0.4625)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0915 (2.2819)  time: 2.0329  data: 0.0098  max mem: 2905
Epoch: [63]  [600/625]  eta: 0:00:51  lr: 0.003762  min_lr: 0.003762  loss: 3.3034 (3.2949)  class_acc: 0.4570 (0.4615)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7658 (2.2952)  time: 2.1978  data: 0.0008  max mem: 2905
Epoch: [63]  [624/625]  eta: 0:00:01  lr: 0.003761  min_lr: 0.003761  loss: 3.3163 (3.2961)  class_acc: 0.4570 (0.4613)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4834 (2.3165)  time: 0.7203  data: 0.0016  max mem: 2905
Epoch: [63] Total time: 0:20:48 (1.9969 s / it)
Averaged stats: lr: 0.003761  min_lr: 0.003761  loss: 3.3163 (3.2990)  class_acc: 0.4570 (0.4616)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4834 (2.3165)
Test:  [ 0/50]  eta: 0:10:58  loss: 2.8210 (2.8210)  acc1: 39.2000 (39.2000)  acc5: 64.8000 (64.8000)  time: 13.1770  data: 13.1492  max mem: 2905
Test:  [10/50]  eta: 0:01:26  loss: 2.5112 (2.5639)  acc1: 49.6000 (47.7818)  acc5: 69.6000 (69.8182)  time: 2.1529  data: 2.1330  max mem: 2905
Test:  [20/50]  eta: 0:00:50  loss: 2.7240 (2.7259)  acc1: 41.6000 (44.1905)  acc5: 67.2000 (67.3143)  time: 1.1248  data: 1.1057  max mem: 2905
Test:  [30/50]  eta: 0:00:30  loss: 2.8464 (2.7408)  acc1: 40.0000 (43.3290)  acc5: 64.8000 (67.5355)  time: 1.1552  data: 1.1357  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 2.7646 (2.7504)  acc1: 40.0000 (42.9268)  acc5: 64.8000 (67.5512)  time: 0.8427  data: 0.8226  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.6743 (2.7355)  acc1: 41.6000 (42.8480)  acc5: 68.0000 (67.9200)  time: 0.7518  data: 0.7322  max mem: 2905
Test: Total time: 0:00:53 (1.0701 s / it)
* Acc@1 43.300 Acc@5 68.072 loss 2.707
Accuracy of the model on the 50000 test images: 43.3%
Max accuracy: 50.54%
Epoch: [64]  [  0/625]  eta: 4:13:30  lr: 0.003761  min_lr: 0.003761  loss: 3.2595 (3.2595)  class_acc: 0.4648 (0.4648)  weight_decay: 0.0500 (0.0500)  time: 24.3364  data: 21.6845  max mem: 2905
Epoch: [64]  [200/625]  eta: 0:14:48  lr: 0.003758  min_lr: 0.003758  loss: 3.3159 (3.2854)  class_acc: 0.4531 (0.4652)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.4851 (2.6517)  time: 1.8460  data: 0.0233  max mem: 2905
Epoch: [64]  [400/625]  eta: 0:07:34  lr: 0.003754  min_lr: 0.003754  loss: 3.2846 (3.2904)  class_acc: 0.4688 (0.4646)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9850 (2.5268)  time: 2.0707  data: 0.0432  max mem: 2905
Epoch: [64]  [600/625]  eta: 0:00:49  lr: 0.003751  min_lr: 0.003751  loss: 3.2551 (3.2931)  class_acc: 0.4648 (0.4638)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.5453 (2.4160)  time: 1.7913  data: 0.0008  max mem: 2905
Epoch: [64]  [624/625]  eta: 0:00:01  lr: 0.003751  min_lr: 0.003751  loss: 3.2928 (3.2933)  class_acc: 0.4453 (0.4638)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8197 (2.4034)  time: 1.1334  data: 0.0013  max mem: 2905
Epoch: [64] Total time: 0:20:16 (1.9471 s / it)
Averaged stats: lr: 0.003751  min_lr: 0.003751  loss: 3.2928 (3.2943)  class_acc: 0.4453 (0.4625)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8197 (2.4034)
Test:  [ 0/50]  eta: 0:10:20  loss: 2.5753 (2.5753)  acc1: 43.2000 (43.2000)  acc5: 73.6000 (73.6000)  time: 12.4140  data: 12.3846  max mem: 2905
Test:  [10/50]  eta: 0:01:26  loss: 2.2326 (2.2686)  acc1: 51.2000 (51.7818)  acc5: 73.6000 (74.6909)  time: 2.1583  data: 2.1376  max mem: 2905
Test:  [20/50]  eta: 0:00:51  loss: 2.4573 (2.4576)  acc1: 45.6000 (47.9238)  acc5: 71.2000 (71.8095)  time: 1.1754  data: 1.1558  max mem: 2905
Test:  [30/50]  eta: 0:00:29  loss: 2.4989 (2.4497)  acc1: 45.6000 (47.8710)  acc5: 69.6000 (72.2839)  time: 1.0760  data: 1.0568  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 2.4338 (2.4724)  acc1: 47.2000 (47.4927)  acc5: 73.6000 (71.8634)  time: 0.6565  data: 0.6360  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 2.4733 (2.4769)  acc1: 47.2000 (47.5520)  acc5: 72.0000 (72.0000)  time: 0.5592  data: 0.5377  max mem: 2905
Test: Total time: 0:00:49 (0.9991 s / it)
* Acc@1 47.702 Acc@5 72.402 loss 2.465
Accuracy of the model on the 50000 test images: 47.7%
Max accuracy: 50.54%
Epoch: [65]  [  0/625]  eta: 3:28:46  lr: 0.003751  min_lr: 0.003751  loss: 3.0546 (3.0546)  class_acc: 0.5234 (0.5234)  weight_decay: 0.0500 (0.0500)  time: 20.0417  data: 17.8946  max mem: 2905
Epoch: [65]  [200/625]  eta: 0:13:52  lr: 0.003747  min_lr: 0.003747  loss: 3.2824 (3.2756)  class_acc: 0.4492 (0.4665)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0486 (2.1572)  time: 1.7947  data: 0.0153  max mem: 2905
Epoch: [65]  [400/625]  eta: 0:07:15  lr: 0.003744  min_lr: 0.003744  loss: 3.2794 (3.2817)  class_acc: 0.4570 (0.4666)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1750 (2.3234)  time: 1.7721  data: 0.0006  max mem: 2905
Epoch: [65]  [600/625]  eta: 0:00:48  lr: 0.003740  min_lr: 0.003740  loss: 3.2455 (3.2915)  class_acc: 0.4648 (0.4642)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1300 (2.3734)  time: 1.8006  data: 0.0007  max mem: 2905
Epoch: [65]  [624/625]  eta: 0:00:01  lr: 0.003740  min_lr: 0.003740  loss: 3.2493 (3.2904)  class_acc: 0.4570 (0.4640)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.5518 (2.3458)  time: 0.5022  data: 0.0013  max mem: 2905
Epoch: [65] Total time: 0:19:46 (1.8983 s / it)
Averaged stats: lr: 0.003740  min_lr: 0.003740  loss: 3.2493 (3.2930)  class_acc: 0.4570 (0.4627)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.5518 (2.3458)
Test:  [ 0/50]  eta: 0:11:23  loss: 3.3565 (3.3565)  acc1: 30.4000 (30.4000)  acc5: 58.4000 (58.4000)  time: 13.6671  data: 13.6356  max mem: 2905
Test:  [10/50]  eta: 0:01:28  loss: 2.6612 (2.7729)  acc1: 45.6000 (43.3455)  acc5: 68.0000 (66.6182)  time: 2.2197  data: 2.1993  max mem: 2905
Test:  [20/50]  eta: 0:00:51  loss: 2.7806 (2.9154)  acc1: 40.8000 (39.4667)  acc5: 66.4000 (65.0286)  time: 1.1249  data: 1.1030  max mem: 2905
Test:  [30/50]  eta: 0:00:30  loss: 2.9907 (2.9645)  acc1: 35.2000 (38.9419)  acc5: 62.4000 (64.2839)  time: 1.1072  data: 1.0860  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 2.9907 (2.9921)  acc1: 37.6000 (38.8488)  acc5: 62.4000 (63.9220)  time: 0.7560  data: 0.7372  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 3.0347 (2.9969)  acc1: 38.4000 (39.0080)  acc5: 63.2000 (63.8560)  time: 0.7157  data: 0.6959  max mem: 2905
Test: Total time: 0:00:51 (1.0373 s / it)
* Acc@1 38.858 Acc@5 64.376 loss 2.958
Accuracy of the model on the 50000 test images: 38.9%
Max accuracy: 50.54%
Epoch: [66]  [  0/625]  eta: 4:13:42  lr: 0.003740  min_lr: 0.003740  loss: 3.1317 (3.1317)  class_acc: 0.5000 (0.5000)  weight_decay: 0.0500 (0.0500)  time: 24.3558  data: 18.3498  max mem: 2905
Epoch: [66]  [200/625]  eta: 0:14:38  lr: 0.003736  min_lr: 0.003736  loss: 3.2005 (3.2693)  class_acc: 0.4883 (0.4694)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.6101 (2.0653)  time: 1.9880  data: 0.0007  max mem: 2905
Epoch: [66]  [400/625]  eta: 0:07:36  lr: 0.003732  min_lr: 0.003732  loss: 3.2568 (3.2826)  class_acc: 0.4727 (0.4666)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8743 (2.2674)  time: 1.9582  data: 0.0007  max mem: 2905
Epoch: [66]  [600/625]  eta: 0:00:50  lr: 0.003729  min_lr: 0.003729  loss: 3.2907 (3.2877)  class_acc: 0.4688 (0.4656)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.8982 (2.3220)  time: 1.9225  data: 0.0008  max mem: 2905
Epoch: [66]  [624/625]  eta: 0:00:01  lr: 0.003728  min_lr: 0.003728  loss: 3.2886 (3.2894)  class_acc: 0.4531 (0.4651)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8921 (2.3150)  time: 0.8207  data: 0.0017  max mem: 2905
Epoch: [66] Total time: 0:20:28 (1.9652 s / it)
Averaged stats: lr: 0.003728  min_lr: 0.003728  loss: 3.2886 (3.2860)  class_acc: 0.4531 (0.4647)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8921 (2.3150)
Test:  [ 0/50]  eta: 0:10:32  loss: 2.5922 (2.5922)  acc1: 45.6000 (45.6000)  acc5: 74.4000 (74.4000)  time: 12.6414  data: 12.6096  max mem: 2905
Test:  [10/50]  eta: 0:01:26  loss: 2.4585 (2.3893)  acc1: 48.0000 (49.5273)  acc5: 74.4000 (73.9636)  time: 2.1562  data: 2.1346  max mem: 2905
Test:  [20/50]  eta: 0:00:51  loss: 2.5236 (2.5445)  acc1: 45.6000 (45.6762)  acc5: 71.2000 (71.4286)  time: 1.1626  data: 1.1431  max mem: 2905
Test:  [30/50]  eta: 0:00:30  loss: 2.5935 (2.5335)  acc1: 42.4000 (45.6258)  acc5: 70.4000 (71.2000)  time: 1.1773  data: 1.1582  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 2.4917 (2.5227)  acc1: 46.4000 (46.1659)  acc5: 71.2000 (71.0439)  time: 0.8395  data: 0.8202  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.5178 (2.5222)  acc1: 46.4000 (46.0960)  acc5: 71.2000 (71.1840)  time: 0.7331  data: 0.7132  max mem: 2905
Test: Total time: 0:00:53 (1.0723 s / it)
* Acc@1 46.272 Acc@5 71.408 loss 2.506
Accuracy of the model on the 50000 test images: 46.3%
Max accuracy: 50.54%
Epoch: [67]  [  0/625]  eta: 3:35:29  lr: 0.003728  min_lr: 0.003728  loss: 3.3032 (3.3032)  class_acc: 0.4453 (0.4453)  weight_decay: 0.0500 (0.0500)  time: 20.6867  data: 16.2919  max mem: 2905
Epoch: [67]  [200/625]  eta: 0:14:11  lr: 0.003725  min_lr: 0.003725  loss: 3.2558 (3.2759)  class_acc: 0.4531 (0.4656)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8062 (2.4144)  time: 1.9989  data: 0.0008  max mem: 2905
Epoch: [67]  [400/625]  eta: 0:07:21  lr: 0.003721  min_lr: 0.003721  loss: 3.2501 (3.2790)  class_acc: 0.4648 (0.4651)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6835 (inf)  time: 1.8869  data: 0.0009  max mem: 2905
Epoch: [67]  [600/625]  eta: 0:00:49  lr: 0.003717  min_lr: 0.003717  loss: 3.3442 (3.2892)  class_acc: 0.4531 (0.4630)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3558 (inf)  time: 1.9344  data: 0.0008  max mem: 2905
Epoch: [67]  [624/625]  eta: 0:00:01  lr: 0.003717  min_lr: 0.003717  loss: 3.3001 (3.2896)  class_acc: 0.4609 (0.4629)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4876 (inf)  time: 0.4860  data: 0.0027  max mem: 2905
Epoch: [67] Total time: 0:20:17 (1.9482 s / it)
Averaged stats: lr: 0.003717  min_lr: 0.003717  loss: 3.3001 (3.2834)  class_acc: 0.4609 (0.4649)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4876 (inf)
Test:  [ 0/50]  eta: 0:10:11  loss: 2.4155 (2.4155)  acc1: 49.6000 (49.6000)  acc5: 75.2000 (75.2000)  time: 12.2247  data: 12.1922  max mem: 2905
Test:  [10/50]  eta: 0:01:23  loss: 2.3761 (2.3209)  acc1: 51.2000 (51.3455)  acc5: 72.8000 (74.9818)  time: 2.0864  data: 2.0654  max mem: 2905
Test:  [20/50]  eta: 0:00:50  loss: 2.5671 (2.5084)  acc1: 46.4000 (47.2381)  acc5: 70.4000 (72.2667)  time: 1.1415  data: 1.1221  max mem: 2905
Test:  [30/50]  eta: 0:00:29  loss: 2.7034 (2.5077)  acc1: 43.2000 (46.8903)  acc5: 69.6000 (71.6387)  time: 1.1633  data: 1.1441  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 2.5463 (2.5183)  acc1: 45.6000 (46.6537)  acc5: 69.6000 (71.3561)  time: 0.8611  data: 0.8416  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.4716 (2.5196)  acc1: 45.6000 (46.4960)  acc5: 68.8000 (71.1040)  time: 0.7859  data: 0.7671  max mem: 2905
Test: Total time: 0:00:52 (1.0524 s / it)
* Acc@1 46.630 Acc@5 71.640 loss 2.499
Accuracy of the model on the 50000 test images: 46.6%
Max accuracy: 50.54%
Epoch: [68]  [  0/625]  eta: 3:35:23  lr: 0.003717  min_lr: 0.003717  loss: 3.3656 (3.3656)  class_acc: 0.4492 (0.4492)  weight_decay: 0.0500 (0.0500)  time: 20.6781  data: 17.5342  max mem: 2905
Epoch: [68]  [200/625]  eta: 0:14:32  lr: 0.003713  min_lr: 0.003713  loss: 3.2633 (3.2471)  class_acc: 0.4648 (0.4744)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9967 (2.2418)  time: 2.0033  data: 0.0008  max mem: 2905
Epoch: [68]  [400/625]  eta: 0:07:29  lr: 0.003710  min_lr: 0.003710  loss: 3.3248 (3.2670)  class_acc: 0.4609 (0.4687)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8128 (2.3008)  time: 1.9474  data: 0.0017  max mem: 2905
Epoch: [68]  [600/625]  eta: 0:00:49  lr: 0.003706  min_lr: 0.003706  loss: 3.2695 (3.2754)  class_acc: 0.4648 (0.4669)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1800 (2.2377)  time: 2.0684  data: 0.0064  max mem: 2905
Epoch: [68]  [624/625]  eta: 0:00:01  lr: 0.003705  min_lr: 0.003705  loss: 3.2841 (3.2756)  class_acc: 0.4453 (0.4665)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5654 (2.2713)  time: 0.8496  data: 0.0015  max mem: 2905
Epoch: [68] Total time: 0:20:19 (1.9518 s / it)
Averaged stats: lr: 0.003705  min_lr: 0.003705  loss: 3.2841 (3.2779)  class_acc: 0.4453 (0.4654)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5654 (2.2713)
Test:  [ 0/50]  eta: 0:10:57  loss: 2.4177 (2.4177)  acc1: 44.0000 (44.0000)  acc5: 72.0000 (72.0000)  time: 13.1497  data: 13.1174  max mem: 2905
Test:  [10/50]  eta: 0:01:32  loss: 2.4550 (2.4474)  acc1: 46.4000 (47.5636)  acc5: 72.0000 (71.7818)  time: 2.3061  data: 2.2873  max mem: 2905
Test:  [20/50]  eta: 0:00:55  loss: 2.5473 (2.5723)  acc1: 44.8000 (45.2952)  acc5: 68.0000 (70.0191)  time: 1.2749  data: 1.2568  max mem: 2905
Test:  [30/50]  eta: 0:00:32  loss: 2.6019 (2.5602)  acc1: 44.0000 (46.1419)  acc5: 67.2000 (70.1677)  time: 1.2130  data: 1.1923  max mem: 2905
Test:  [40/50]  eta: 0:00:13  loss: 2.6563 (2.6148)  acc1: 45.6000 (46.0098)  acc5: 68.0000 (69.5024)  time: 0.7932  data: 0.7721  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.7203 (2.6443)  acc1: 44.0000 (45.5840)  acc5: 68.8000 (69.3760)  time: 0.7319  data: 0.7132  max mem: 2905
Test: Total time: 0:00:55 (1.1041 s / it)
* Acc@1 44.886 Acc@5 69.590 loss 2.635
Accuracy of the model on the 50000 test images: 44.9%
Max accuracy: 50.54%
Epoch: [69]  [  0/625]  eta: 3:47:55  lr: 0.003705  min_lr: 0.003705  loss: 3.2073 (3.2073)  class_acc: 0.5039 (0.5039)  weight_decay: 0.0500 (0.0500)  time: 21.8809  data: 18.4495  max mem: 2905
Epoch: [69]  [200/625]  eta: 0:14:43  lr: 0.003702  min_lr: 0.003702  loss: 3.2782 (3.2626)  class_acc: 0.4688 (0.4683)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8088 (2.2343)  time: 2.0267  data: 0.1072  max mem: 2905
Epoch: [69]  [400/625]  eta: 0:07:30  lr: 0.003698  min_lr: 0.003698  loss: 3.2588 (3.2713)  class_acc: 0.4648 (0.4668)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8190 (2.2371)  time: 1.7620  data: 0.0054  max mem: 2905
Epoch: [69]  [600/625]  eta: 0:00:50  lr: 0.003694  min_lr: 0.003694  loss: 3.2714 (3.2814)  class_acc: 0.4648 (0.4644)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9751 (2.3004)  time: 2.0135  data: 0.0411  max mem: 2905
Epoch: [69]  [624/625]  eta: 0:00:01  lr: 0.003694  min_lr: 0.003694  loss: 3.2842 (3.2822)  class_acc: 0.4570 (0.4643)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9751 (2.3004)  time: 0.7078  data: 0.0012  max mem: 2905
Epoch: [69] Total time: 0:20:18 (1.9504 s / it)
Averaged stats: lr: 0.003694  min_lr: 0.003694  loss: 3.2842 (3.2774)  class_acc: 0.4570 (0.4661)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9751 (2.3004)
Test:  [ 0/50]  eta: 0:10:14  loss: 2.5352 (2.5352)  acc1: 47.2000 (47.2000)  acc5: 72.8000 (72.8000)  time: 12.2867  data: 12.2539  max mem: 2905
Test:  [10/50]  eta: 0:01:21  loss: 2.3039 (2.2981)  acc1: 50.4000 (50.6909)  acc5: 73.6000 (74.0364)  time: 2.0250  data: 2.0035  max mem: 2905
Test:  [20/50]  eta: 0:00:47  loss: 2.4613 (2.4317)  acc1: 48.0000 (48.0762)  acc5: 72.8000 (72.5714)  time: 1.0466  data: 1.0250  max mem: 2905
Test:  [30/50]  eta: 0:00:26  loss: 2.5963 (2.4802)  acc1: 44.8000 (47.2516)  acc5: 70.4000 (71.4323)  time: 0.9370  data: 0.9161  max mem: 2905
Test:  [40/50]  eta: 0:00:10  loss: 2.5883 (2.4845)  acc1: 44.0000 (47.1024)  acc5: 70.4000 (71.1805)  time: 0.5641  data: 0.5457  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 2.5888 (2.5088)  acc1: 43.2000 (46.5440)  acc5: 69.6000 (70.8960)  time: 0.5196  data: 0.5018  max mem: 2905
Test: Total time: 0:00:45 (0.9007 s / it)
* Acc@1 46.142 Acc@5 71.584 loss 2.487
Accuracy of the model on the 50000 test images: 46.1%
Max accuracy: 50.54%
Epoch: [70]  [  0/625]  eta: 3:31:54  lr: 0.003694  min_lr: 0.003694  loss: 3.1680 (3.1680)  class_acc: 0.4961 (0.4961)  weight_decay: 0.0500 (0.0500)  time: 20.3439  data: 19.5478  max mem: 2905
Epoch: [70]  [200/625]  eta: 0:14:13  lr: 0.003690  min_lr: 0.003690  loss: 3.2342 (3.2535)  class_acc: 0.4688 (0.4697)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8662 (2.0418)  time: 1.8954  data: 1.1768  max mem: 2905
Epoch: [70]  [400/625]  eta: 0:07:21  lr: 0.003686  min_lr: 0.003686  loss: 3.2927 (3.2590)  class_acc: 0.4609 (0.4698)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2044 (2.1714)  time: 1.9010  data: 1.6722  max mem: 2905
Epoch: [70]  [600/625]  eta: 0:00:48  lr: 0.003682  min_lr: 0.003682  loss: 3.2808 (3.2661)  class_acc: 0.4648 (0.4688)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2191 (2.2436)  time: 1.8598  data: 1.6945  max mem: 2905
Epoch: [70]  [624/625]  eta: 0:00:01  lr: 0.003682  min_lr: 0.003682  loss: 3.2867 (3.2673)  class_acc: 0.4531 (0.4683)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.5888 (2.2310)  time: 0.8394  data: 0.6855  max mem: 2905
Epoch: [70] Total time: 0:19:45 (1.8970 s / it)
Averaged stats: lr: 0.003682  min_lr: 0.003682  loss: 3.2867 (3.2710)  class_acc: 0.4531 (0.4675)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.5888 (2.2310)
Test:  [ 0/50]  eta: 0:09:43  loss: 2.3033 (2.3033)  acc1: 44.8000 (44.8000)  acc5: 75.2000 (75.2000)  time: 11.6675  data: 11.6301  max mem: 2905
Test:  [10/50]  eta: 0:01:23  loss: 2.4451 (2.5108)  acc1: 47.2000 (47.2727)  acc5: 71.2000 (71.4909)  time: 2.0781  data: 2.0551  max mem: 2905
Test:  [20/50]  eta: 0:00:49  loss: 2.5803 (2.6281)  acc1: 45.6000 (44.8381)  acc5: 69.6000 (70.1714)  time: 1.1449  data: 1.1249  max mem: 2905
Test:  [30/50]  eta: 0:00:26  loss: 2.6033 (2.6347)  acc1: 40.0000 (44.6194)  acc5: 69.6000 (70.4774)  time: 0.9306  data: 0.9109  max mem: 2905
Test:  [40/50]  eta: 0:00:10  loss: 2.6644 (2.6802)  acc1: 44.8000 (44.1366)  acc5: 68.0000 (69.4829)  time: 0.4902  data: 0.4692  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 2.7450 (2.6783)  acc1: 41.6000 (44.2240)  acc5: 67.2000 (69.4720)  time: 0.4627  data: 0.4426  max mem: 2905
Test: Total time: 0:00:45 (0.9005 s / it)
* Acc@1 44.674 Acc@5 69.370 loss 2.661
Accuracy of the model on the 50000 test images: 44.7%
Max accuracy: 50.54%
Epoch: [71]  [  0/625]  eta: 3:58:25  lr: 0.003681  min_lr: 0.003681  loss: 3.3352 (3.3352)  class_acc: 0.4648 (0.4648)  weight_decay: 0.0500 (0.0500)  time: 22.8884  data: 20.4500  max mem: 2905
Epoch: [71]  [200/625]  eta: 0:13:31  lr: 0.003678  min_lr: 0.003678  loss: 3.2586 (3.2437)  class_acc: 0.4609 (0.4738)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2543 (2.4048)  time: 1.8935  data: 0.0062  max mem: 2905
Epoch: [71]  [400/625]  eta: 0:07:12  lr: 0.003674  min_lr: 0.003674  loss: 3.1917 (3.2518)  class_acc: 0.4844 (0.4710)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3252 (2.2583)  time: 1.9401  data: 0.0230  max mem: 2905
Epoch: [71]  [600/625]  eta: 0:00:48  lr: 0.003670  min_lr: 0.003670  loss: 3.2620 (3.2606)  class_acc: 0.4570 (0.4691)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.6146 (2.2707)  time: 1.8943  data: 0.0006  max mem: 2905
Epoch: [71]  [624/625]  eta: 0:00:01  lr: 0.003669  min_lr: 0.003669  loss: 3.2440 (3.2606)  class_acc: 0.4766 (0.4692)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9048 (2.2708)  time: 0.3730  data: 0.0108  max mem: 2905
Epoch: [71] Total time: 0:19:59 (1.9188 s / it)
Averaged stats: lr: 0.003669  min_lr: 0.003669  loss: 3.2440 (3.2653)  class_acc: 0.4766 (0.4688)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9048 (2.2708)
Test:  [ 0/50]  eta: 0:10:34  loss: 2.8029 (2.8029)  acc1: 43.2000 (43.2000)  acc5: 64.0000 (64.0000)  time: 12.6890  data: 12.6506  max mem: 2905
Test:  [10/50]  eta: 0:01:22  loss: 2.5863 (2.6212)  acc1: 45.6000 (46.2545)  acc5: 69.6000 (69.9636)  time: 2.0631  data: 2.0422  max mem: 2905
Test:  [20/50]  eta: 0:00:48  loss: 2.8255 (2.7735)  acc1: 40.8000 (42.1333)  acc5: 68.0000 (67.3905)  time: 1.0703  data: 1.0511  max mem: 2905
Test:  [30/50]  eta: 0:00:29  loss: 2.8536 (2.7821)  acc1: 38.4000 (41.6774)  acc5: 66.4000 (67.2000)  time: 1.1133  data: 1.0934  max mem: 2905
Test:  [40/50]  eta: 0:00:13  loss: 2.8018 (2.8022)  acc1: 42.4000 (41.6585)  acc5: 67.2000 (67.0829)  time: 1.0214  data: 1.0019  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.7924 (2.7989)  acc1: 39.2000 (41.6000)  acc5: 68.0000 (67.1360)  time: 0.9269  data: 0.9072  max mem: 2905
Test: Total time: 0:00:57 (1.1533 s / it)
* Acc@1 42.048 Acc@5 67.582 loss 2.781
Accuracy of the model on the 50000 test images: 42.0%
Max accuracy: 50.54%
Epoch: [72]  [  0/625]  eta: 4:45:59  lr: 0.003669  min_lr: 0.003669  loss: 3.2441 (3.2441)  class_acc: 0.4609 (0.4609)  weight_decay: 0.0500 (0.0500)  time: 27.4554  data: 27.3179  max mem: 2905
Epoch: [72]  [200/625]  eta: 0:14:47  lr: 0.003665  min_lr: 0.003665  loss: 3.2518 (3.2295)  class_acc: 0.4688 (0.4770)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1179 (2.3574)  time: 1.9984  data: 0.6676  max mem: 2905
Epoch: [72]  [400/625]  eta: 0:07:29  lr: 0.003661  min_lr: 0.003661  loss: 3.2914 (3.2553)  class_acc: 0.4609 (0.4712)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7572 (2.3537)  time: 1.9639  data: 0.0008  max mem: 2905
Epoch: [72]  [600/625]  eta: 0:00:50  lr: 0.003657  min_lr: 0.003657  loss: 3.2522 (3.2603)  class_acc: 0.4609 (0.4706)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.5007 (2.3501)  time: 2.0197  data: 0.0009  max mem: 2905
Epoch: [72]  [624/625]  eta: 0:00:01  lr: 0.003657  min_lr: 0.003657  loss: 3.2724 (3.2622)  class_acc: 0.4609 (0.4700)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0141 (2.3422)  time: 0.7504  data: 0.0107  max mem: 2905
Epoch: [72] Total time: 0:20:19 (1.9520 s / it)
Averaged stats: lr: 0.003657  min_lr: 0.003657  loss: 3.2724 (3.2626)  class_acc: 0.4609 (0.4694)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0141 (2.3422)
Test:  [ 0/50]  eta: 0:10:50  loss: 2.9999 (2.9999)  acc1: 39.2000 (39.2000)  acc5: 68.0000 (68.0000)  time: 13.0025  data: 12.9768  max mem: 2905
Test:  [10/50]  eta: 0:01:20  loss: 2.4491 (2.4742)  acc1: 49.6000 (47.4909)  acc5: 72.0000 (71.4909)  time: 2.0189  data: 1.9978  max mem: 2905
Test:  [20/50]  eta: 0:00:46  loss: 2.5583 (2.6920)  acc1: 40.0000 (42.8190)  acc5: 68.8000 (68.5333)  time: 0.9823  data: 0.9625  max mem: 2905
Test:  [30/50]  eta: 0:00:26  loss: 2.8917 (2.7317)  acc1: 39.2000 (42.7097)  acc5: 64.8000 (67.6387)  time: 0.9727  data: 0.9534  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 2.8917 (2.7469)  acc1: 40.8000 (42.7902)  acc5: 64.8000 (67.4732)  time: 0.6816  data: 0.6608  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 2.8702 (2.7662)  acc1: 40.8000 (42.4800)  acc5: 65.6000 (67.5520)  time: 0.5656  data: 0.5451  max mem: 2905
Test: Total time: 0:00:48 (0.9613 s / it)
* Acc@1 42.372 Acc@5 67.640 loss 2.754
Accuracy of the model on the 50000 test images: 42.4%
Max accuracy: 50.54%
Epoch: [73]  [  0/625]  eta: 3:41:34  lr: 0.003657  min_lr: 0.003657  loss: 3.3996 (3.3996)  class_acc: 0.4336 (0.4336)  weight_decay: 0.0500 (0.0500)  time: 21.2717  data: 21.1469  max mem: 2905
Epoch: [73]  [200/625]  eta: 0:14:18  lr: 0.003653  min_lr: 0.003653  loss: 3.2344 (3.2404)  class_acc: 0.4648 (0.4727)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8714 (2.3847)  time: 2.1308  data: 0.2239  max mem: 2905
Epoch: [73]  [400/625]  eta: 0:07:24  lr: 0.003649  min_lr: 0.003649  loss: 3.2630 (3.2619)  class_acc: 0.4609 (0.4695)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.5318 (2.4013)  time: 2.0914  data: 0.4550  max mem: 2905
Epoch: [73]  [600/625]  eta: 0:00:49  lr: 0.003645  min_lr: 0.003645  loss: 3.1777 (3.2662)  class_acc: 0.4688 (0.4683)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.5482 (2.4028)  time: 1.9119  data: 0.0007  max mem: 2905
Epoch: [73]  [624/625]  eta: 0:00:01  lr: 0.003644  min_lr: 0.003644  loss: 3.2238 (3.2663)  class_acc: 0.4688 (0.4683)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1276 (2.4070)  time: 0.6894  data: 0.0195  max mem: 2905
Epoch: [73] Total time: 0:20:13 (1.9422 s / it)
Averaged stats: lr: 0.003644  min_lr: 0.003644  loss: 3.2238 (3.2615)  class_acc: 0.4688 (0.4691)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1276 (2.4070)
Test:  [ 0/50]  eta: 0:10:19  loss: 2.4044 (2.4044)  acc1: 48.0000 (48.0000)  acc5: 73.6000 (73.6000)  time: 12.3961  data: 12.3669  max mem: 2905
Test:  [10/50]  eta: 0:01:23  loss: 2.4581 (2.4322)  acc1: 48.0000 (48.6545)  acc5: 73.6000 (73.6727)  time: 2.0979  data: 2.0774  max mem: 2905
Test:  [20/50]  eta: 0:00:49  loss: 2.5647 (2.5454)  acc1: 44.0000 (45.7524)  acc5: 72.0000 (71.5429)  time: 1.1155  data: 1.0947  max mem: 2905
Test:  [30/50]  eta: 0:00:28  loss: 2.5929 (2.5454)  acc1: 41.6000 (45.3419)  acc5: 70.4000 (70.8129)  time: 1.0391  data: 1.0181  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 2.5929 (2.6031)  acc1: 41.6000 (44.2146)  acc5: 68.8000 (70.1854)  time: 0.6100  data: 0.5895  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 2.7911 (2.6439)  acc1: 40.8000 (43.6000)  acc5: 68.8000 (69.7120)  time: 0.5337  data: 0.5132  max mem: 2905
Test: Total time: 0:00:47 (0.9540 s / it)
* Acc@1 44.012 Acc@5 69.468 loss 2.622
Accuracy of the model on the 50000 test images: 44.0%
Max accuracy: 50.54%
Epoch: [74]  [  0/625]  eta: 3:22:34  lr: 0.003644  min_lr: 0.003644  loss: 3.2778 (3.2778)  class_acc: 0.4297 (0.4297)  weight_decay: 0.0500 (0.0500)  time: 19.4468  data: 17.8797  max mem: 2905
Epoch: [74]  [200/625]  eta: 0:13:44  lr: 0.003640  min_lr: 0.003640  loss: 3.2522 (3.2521)  class_acc: 0.4805 (0.4716)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9541 (inf)  time: 1.7463  data: 0.9433  max mem: 2905
Epoch: [74]  [400/625]  eta: 0:06:59  lr: 0.003636  min_lr: 0.003636  loss: 3.2739 (3.2589)  class_acc: 0.4688 (0.4695)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7812 (inf)  time: 1.7296  data: 1.4300  max mem: 2905
Epoch: [74]  [600/625]  eta: 0:00:46  lr: 0.003632  min_lr: 0.003632  loss: 3.2757 (3.2619)  class_acc: 0.4531 (0.4688)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6342 (inf)  time: 1.8961  data: 1.7076  max mem: 2905
Epoch: [74]  [624/625]  eta: 0:00:01  lr: 0.003631  min_lr: 0.003631  loss: 3.2375 (3.2617)  class_acc: 0.4688 (0.4686)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1911 (inf)  time: 0.7458  data: 0.5926  max mem: 2905
Epoch: [74] Total time: 0:19:14 (1.8473 s / it)
Averaged stats: lr: 0.003631  min_lr: 0.003631  loss: 3.2375 (3.2609)  class_acc: 0.4688 (0.4693)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1911 (inf)
Test:  [ 0/50]  eta: 0:10:40  loss: 2.7115 (2.7115)  acc1: 40.0000 (40.0000)  acc5: 67.2000 (67.2000)  time: 12.8154  data: 12.7897  max mem: 2905
Test:  [10/50]  eta: 0:01:21  loss: 2.5669 (2.5203)  acc1: 48.0000 (48.1455)  acc5: 69.6000 (71.2727)  time: 2.0346  data: 2.0141  max mem: 2905
Test:  [20/50]  eta: 0:00:46  loss: 2.5740 (2.6284)  acc1: 47.2000 (45.4857)  acc5: 69.6000 (69.9429)  time: 0.9883  data: 0.9675  max mem: 2905
Test:  [30/50]  eta: 0:00:27  loss: 2.5641 (2.6008)  acc1: 43.2000 (45.5226)  acc5: 68.8000 (69.9355)  time: 0.9944  data: 0.9739  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 2.5385 (2.6105)  acc1: 44.8000 (45.5805)  acc5: 68.8000 (69.8732)  time: 0.8006  data: 0.7817  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.5839 (2.6133)  acc1: 47.2000 (45.4240)  acc5: 69.6000 (69.9040)  time: 0.6633  data: 0.6441  max mem: 2905
Test: Total time: 0:00:52 (1.0472 s / it)
* Acc@1 45.582 Acc@5 70.386 loss 2.583
Accuracy of the model on the 50000 test images: 45.6%
Max accuracy: 50.54%
Epoch: [75]  [  0/625]  eta: 3:14:11  lr: 0.003631  min_lr: 0.003631  loss: 3.3014 (3.3014)  class_acc: 0.4688 (0.4688)  weight_decay: 0.0500 (0.0500)  time: 18.6430  data: 17.9810  max mem: 2905
Epoch: [75]  [200/625]  eta: 0:13:31  lr: 0.003627  min_lr: 0.003627  loss: 3.2851 (3.2233)  class_acc: 0.4648 (0.4764)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.4224 (2.2634)  time: 1.9488  data: 1.1112  max mem: 2905
Epoch: [75]  [400/625]  eta: 0:07:00  lr: 0.003623  min_lr: 0.003623  loss: 3.2408 (3.2366)  class_acc: 0.4727 (0.4742)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2368 (2.3292)  time: 1.8082  data: 0.8952  max mem: 2905
Epoch: [75]  [600/625]  eta: 0:00:46  lr: 0.003619  min_lr: 0.003619  loss: 3.2120 (3.2422)  class_acc: 0.4648 (0.4730)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9178 (2.2967)  time: 1.7674  data: 1.3425  max mem: 2905
Epoch: [75]  [624/625]  eta: 0:00:01  lr: 0.003618  min_lr: 0.003618  loss: 3.2393 (3.2418)  class_acc: 0.4648 (0.4730)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8435 (2.3120)  time: 0.6558  data: 0.4252  max mem: 2905
Epoch: [75] Total time: 0:19:18 (1.8528 s / it)
Averaged stats: lr: 0.003618  min_lr: 0.003618  loss: 3.2393 (3.2509)  class_acc: 0.4648 (0.4714)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8435 (2.3120)
Test:  [ 0/50]  eta: 0:09:24  loss: 3.1796 (3.1796)  acc1: 31.2000 (31.2000)  acc5: 55.2000 (55.2000)  time: 11.2840  data: 11.2471  max mem: 2905
Test:  [10/50]  eta: 0:01:15  loss: 2.8774 (2.8887)  acc1: 42.4000 (39.7091)  acc5: 65.6000 (65.7455)  time: 1.8980  data: 1.8789  max mem: 2905
Test:  [20/50]  eta: 0:00:45  loss: 2.9550 (2.9869)  acc1: 36.0000 (37.8286)  acc5: 64.0000 (63.7714)  time: 1.0258  data: 1.0064  max mem: 2905
Test:  [30/50]  eta: 0:00:26  loss: 2.9550 (2.9381)  acc1: 37.6000 (39.2516)  acc5: 64.0000 (64.3355)  time: 1.0165  data: 0.9962  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 2.8604 (2.9443)  acc1: 40.0000 (39.1024)  acc5: 64.0000 (64.5268)  time: 0.7972  data: 0.7755  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 3.0163 (2.9519)  acc1: 37.6000 (39.4400)  acc5: 62.4000 (64.4640)  time: 0.6565  data: 0.6338  max mem: 2905
Test: Total time: 0:00:51 (1.0202 s / it)
* Acc@1 39.926 Acc@5 65.324 loss 2.914
Accuracy of the model on the 50000 test images: 39.9%
Max accuracy: 50.54%
Epoch: [76]  [  0/625]  eta: 3:19:10  lr: 0.003618  min_lr: 0.003618  loss: 3.2141 (3.2141)  class_acc: 0.5000 (0.5000)  weight_decay: 0.0500 (0.0500)  time: 19.1208  data: 16.4793  max mem: 2905
Epoch: [76]  [200/625]  eta: 0:13:34  lr: 0.003614  min_lr: 0.003614  loss: 3.2017 (3.2141)  class_acc: 0.4805 (0.4795)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7372 (2.1841)  time: 1.9545  data: 0.0006  max mem: 2905
Epoch: [76]  [400/625]  eta: 0:07:08  lr: 0.003610  min_lr: 0.003610  loss: 3.2270 (3.2342)  class_acc: 0.4727 (0.4757)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0285 (2.2051)  time: 1.8680  data: 0.0011  max mem: 2905
Epoch: [76]  [600/625]  eta: 0:00:47  lr: 0.003605  min_lr: 0.003605  loss: 3.2251 (3.2480)  class_acc: 0.4805 (0.4729)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9673 (2.2847)  time: 1.9398  data: 0.0009  max mem: 2905
Epoch: [76]  [624/625]  eta: 0:00:01  lr: 0.003605  min_lr: 0.003605  loss: 3.1834 (3.2478)  class_acc: 0.4648 (0.4727)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8908 (2.2698)  time: 0.7790  data: 0.0017  max mem: 2905
Epoch: [76] Total time: 0:19:30 (1.8723 s / it)
Averaged stats: lr: 0.003605  min_lr: 0.003605  loss: 3.1834 (3.2505)  class_acc: 0.4648 (0.4716)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8908 (2.2698)
Test:  [ 0/50]  eta: 0:09:56  loss: 2.8905 (2.8905)  acc1: 40.8000 (40.8000)  acc5: 68.0000 (68.0000)  time: 11.9219  data: 11.8856  max mem: 2905
Test:  [10/50]  eta: 0:01:18  loss: 2.5949 (2.4659)  acc1: 48.0000 (48.3636)  acc5: 71.2000 (73.2364)  time: 1.9684  data: 1.9461  max mem: 2905
Test:  [20/50]  eta: 0:00:45  loss: 2.6713 (2.6873)  acc1: 44.8000 (44.1905)  acc5: 70.4000 (70.1333)  time: 0.9806  data: 0.9604  max mem: 2905
Test:  [30/50]  eta: 0:00:25  loss: 2.7028 (2.6572)  acc1: 42.4000 (44.6194)  acc5: 68.0000 (70.1161)  time: 0.8688  data: 0.8497  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 2.5708 (2.6823)  acc1: 45.6000 (44.3512)  acc5: 68.0000 (69.5805)  time: 0.6970  data: 0.6789  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 2.8549 (2.7078)  acc1: 40.8000 (44.0320)  acc5: 67.2000 (69.3440)  time: 0.4666  data: 0.4481  max mem: 2905
Test: Total time: 0:00:48 (0.9704 s / it)
* Acc@1 44.416 Acc@5 68.872 loss 2.693
Accuracy of the model on the 50000 test images: 44.4%
Max accuracy: 50.54%
Epoch: [77]  [  0/625]  eta: 3:23:02  lr: 0.003605  min_lr: 0.003605  loss: 3.1335 (3.1335)  class_acc: 0.5000 (0.5000)  weight_decay: 0.0500 (0.0500)  time: 19.4922  data: 17.2747  max mem: 2905
Epoch: [77]  [200/625]  eta: 0:14:45  lr: 0.003601  min_lr: 0.003601  loss: 3.2871 (3.2283)  class_acc: 0.4609 (0.4747)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2438 (2.3596)  time: 2.2775  data: 0.0142  max mem: 2905
Epoch: [77]  [400/625]  eta: 0:07:39  lr: 0.003596  min_lr: 0.003596  loss: 3.1919 (3.2328)  class_acc: 0.4766 (0.4741)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8472 (2.4034)  time: 2.0814  data: 0.0009  max mem: 2905
Epoch: [77]  [600/625]  eta: 0:00:51  lr: 0.003592  min_lr: 0.003592  loss: 3.2658 (3.2416)  class_acc: 0.4570 (0.4725)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.4974 (2.3675)  time: 2.0703  data: 0.0008  max mem: 2905
Epoch: [77]  [624/625]  eta: 0:00:01  lr: 0.003591  min_lr: 0.003591  loss: 3.2584 (3.2431)  class_acc: 0.4609 (0.4724)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9946 (2.3511)  time: 0.7879  data: 0.0014  max mem: 2905
Epoch: [77] Total time: 0:20:44 (1.9913 s / it)
Averaged stats: lr: 0.003591  min_lr: 0.003591  loss: 3.2584 (3.2502)  class_acc: 0.4609 (0.4714)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9946 (2.3511)
Test:  [ 0/50]  eta: 0:10:26  loss: 2.8636 (2.8636)  acc1: 39.2000 (39.2000)  acc5: 59.2000 (59.2000)  time: 12.5368  data: 12.5089  max mem: 2905
Test:  [10/50]  eta: 0:01:18  loss: 2.7026 (2.6826)  acc1: 45.6000 (44.9455)  acc5: 68.0000 (67.4909)  time: 1.9725  data: 1.9526  max mem: 2905
Test:  [20/50]  eta: 0:00:45  loss: 2.9940 (2.9806)  acc1: 37.6000 (39.1619)  acc5: 63.2000 (63.7714)  time: 0.9714  data: 0.9530  max mem: 2905
Test:  [30/50]  eta: 0:00:26  loss: 3.1087 (2.9214)  acc1: 35.2000 (39.7161)  acc5: 61.6000 (64.9290)  time: 0.9758  data: 0.9571  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 3.0187 (2.9601)  acc1: 36.0000 (38.5561)  acc5: 64.0000 (64.5854)  time: 0.9415  data: 0.9221  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 3.0477 (2.9741)  acc1: 35.2000 (38.5760)  acc5: 63.2000 (64.5120)  time: 0.7431  data: 0.7246  max mem: 2905
Test: Total time: 0:00:55 (1.1072 s / it)
* Acc@1 39.770 Acc@5 65.256 loss 2.925
Accuracy of the model on the 50000 test images: 39.8%
Max accuracy: 50.54%
Epoch: [78]  [  0/625]  eta: 3:59:48  lr: 0.003591  min_lr: 0.003591  loss: 3.0610 (3.0610)  class_acc: 0.5508 (0.5508)  weight_decay: 0.0500 (0.0500)  time: 23.0222  data: 18.9652  max mem: 2905
Epoch: [78]  [200/625]  eta: 0:14:45  lr: 0.003587  min_lr: 0.003587  loss: 3.2425 (3.2196)  class_acc: 0.4805 (0.4769)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9498 (2.1774)  time: 1.8999  data: 0.0293  max mem: 2905
Epoch: [78]  [400/625]  eta: 0:07:37  lr: 0.003583  min_lr: 0.003583  loss: 3.2560 (3.2415)  class_acc: 0.4766 (0.4729)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7879 (2.3186)  time: 1.9466  data: 0.0009  max mem: 2905
Epoch: [78]  [600/625]  eta: 0:00:50  lr: 0.003578  min_lr: 0.003578  loss: 3.2755 (3.2462)  class_acc: 0.4688 (0.4720)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4751 (2.3692)  time: 2.1851  data: 0.0070  max mem: 2905
Epoch: [78]  [624/625]  eta: 0:00:01  lr: 0.003578  min_lr: 0.003578  loss: 3.2972 (3.2478)  class_acc: 0.4570 (0.4716)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4421 (2.3766)  time: 0.6817  data: 0.0013  max mem: 2905
Epoch: [78] Total time: 0:20:27 (1.9640 s / it)
Averaged stats: lr: 0.003578  min_lr: 0.003578  loss: 3.2972 (3.2454)  class_acc: 0.4570 (0.4725)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4421 (2.3766)
Test:  [ 0/50]  eta: 0:10:01  loss: 2.6390 (2.6390)  acc1: 40.0000 (40.0000)  acc5: 71.2000 (71.2000)  time: 12.0390  data: 12.0108  max mem: 2905
Test:  [10/50]  eta: 0:01:15  loss: 2.5639 (2.5345)  acc1: 46.4000 (46.7636)  acc5: 71.2000 (70.1818)  time: 1.8892  data: 1.8698  max mem: 2905
Test:  [20/50]  eta: 0:00:40  loss: 2.7075 (2.6961)  acc1: 41.6000 (42.9714)  acc5: 68.8000 (69.2190)  time: 0.8176  data: 0.7971  max mem: 2905
Test:  [30/50]  eta: 0:00:22  loss: 2.7879 (2.7247)  acc1: 40.0000 (42.8129)  acc5: 68.8000 (69.0065)  time: 0.6937  data: 0.6735  max mem: 2905
Test:  [40/50]  eta: 0:00:09  loss: 2.6464 (2.7150)  acc1: 41.6000 (43.1805)  acc5: 69.6000 (69.0732)  time: 0.6179  data: 0.5991  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 2.8301 (2.7481)  acc1: 41.6000 (43.0560)  acc5: 66.4000 (68.5920)  time: 0.5397  data: 0.5214  max mem: 2905
Test: Total time: 0:00:44 (0.8888 s / it)
* Acc@1 43.550 Acc@5 68.968 loss 2.711
Accuracy of the model on the 50000 test images: 43.6%
Max accuracy: 50.54%
Epoch: [79]  [  0/625]  eta: 3:36:25  lr: 0.003578  min_lr: 0.003578  loss: 3.2456 (3.2456)  class_acc: 0.4648 (0.4648)  weight_decay: 0.0500 (0.0500)  time: 20.7761  data: 17.3916  max mem: 2905
Epoch: [79]  [200/625]  eta: 0:13:49  lr: 0.003573  min_lr: 0.003573  loss: 3.2224 (3.2172)  class_acc: 0.4648 (0.4782)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2990 (2.4234)  time: 1.8416  data: 0.4426  max mem: 2905
Epoch: [79]  [400/625]  eta: 0:07:13  lr: 0.003569  min_lr: 0.003569  loss: 3.2415 (3.2276)  class_acc: 0.4727 (0.4769)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7812 (2.2080)  time: 1.9107  data: 0.6879  max mem: 2905
Epoch: [79]  [600/625]  eta: 0:00:47  lr: 0.003564  min_lr: 0.003564  loss: 3.2019 (3.2301)  class_acc: 0.4805 (0.4759)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7941 (2.1904)  time: 1.9262  data: 0.0038  max mem: 2905
Epoch: [79]  [624/625]  eta: 0:00:01  lr: 0.003564  min_lr: 0.003564  loss: 3.2196 (3.2311)  class_acc: 0.4727 (0.4756)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3000 (2.2107)  time: 0.6290  data: 0.0178  max mem: 2905
Epoch: [79] Total time: 0:19:36 (1.8820 s / it)
Averaged stats: lr: 0.003564  min_lr: 0.003564  loss: 3.2196 (3.2409)  class_acc: 0.4727 (0.4737)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3000 (2.2107)
Test:  [ 0/50]  eta: 0:11:04  loss: 2.1923 (2.1923)  acc1: 48.0000 (48.0000)  acc5: 80.8000 (80.8000)  time: 13.2909  data: 13.2682  max mem: 2905
Test:  [10/50]  eta: 0:01:26  loss: 2.5050 (2.4486)  acc1: 49.6000 (48.0727)  acc5: 74.4000 (72.1455)  time: 2.1504  data: 2.1306  max mem: 2905
Test:  [20/50]  eta: 0:00:46  loss: 2.6015 (2.6043)  acc1: 46.4000 (45.2952)  acc5: 68.0000 (69.6381)  time: 0.9575  data: 0.9380  max mem: 2905
Test:  [30/50]  eta: 0:00:25  loss: 2.5778 (2.5703)  acc1: 44.8000 (45.7032)  acc5: 68.8000 (70.4258)  time: 0.7868  data: 0.7667  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 2.5778 (2.5978)  acc1: 44.8000 (45.4049)  acc5: 71.2000 (70.0293)  time: 0.6966  data: 0.6760  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 2.6982 (2.6245)  acc1: 43.2000 (44.8800)  acc5: 68.8000 (69.8880)  time: 0.5687  data: 0.5493  max mem: 2905
Test: Total time: 0:00:49 (0.9884 s / it)
* Acc@1 45.212 Acc@5 70.668 loss 2.593
Accuracy of the model on the 50000 test images: 45.2%
Max accuracy: 50.54%
Epoch: [80]  [  0/625]  eta: 3:30:04  lr: 0.003564  min_lr: 0.003564  loss: 3.2491 (3.2491)  class_acc: 0.4961 (0.4961)  weight_decay: 0.0500 (0.0500)  time: 20.1677  data: 17.6115  max mem: 2905
Epoch: [80]  [200/625]  eta: 0:13:41  lr: 0.003559  min_lr: 0.003559  loss: 3.2224 (3.2172)  class_acc: 0.4844 (0.4787)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1317 (2.3878)  time: 1.7900  data: 0.2065  max mem: 2905
Epoch: [80]  [400/625]  eta: 0:07:17  lr: 0.003555  min_lr: 0.003555  loss: 3.2049 (3.2250)  class_acc: 0.4688 (0.4770)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9495 (2.4040)  time: 2.0370  data: 0.0009  max mem: 2905
Epoch: [80]  [600/625]  eta: 0:00:48  lr: 0.003550  min_lr: 0.003550  loss: 3.2427 (3.2333)  class_acc: 0.4805 (0.4759)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1451 (inf)  time: 1.7995  data: 0.0720  max mem: 2905
Epoch: [80]  [624/625]  eta: 0:00:01  lr: 0.003550  min_lr: 0.003550  loss: 3.1767 (3.2322)  class_acc: 0.4688 (0.4761)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7529 (inf)  time: 0.4759  data: 0.0063  max mem: 2905
Epoch: [80] Total time: 0:19:53 (1.9102 s / it)
Averaged stats: lr: 0.003550  min_lr: 0.003550  loss: 3.1767 (3.2392)  class_acc: 0.4688 (0.4742)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7529 (inf)
Test:  [ 0/50]  eta: 0:11:01  loss: 3.0755 (3.0755)  acc1: 34.4000 (34.4000)  acc5: 60.8000 (60.8000)  time: 13.2301  data: 13.1962  max mem: 2905
Test:  [10/50]  eta: 0:01:21  loss: 2.9662 (3.0095)  acc1: 40.0000 (39.4909)  acc5: 63.2000 (62.9818)  time: 2.0406  data: 2.0188  max mem: 2905
Test:  [20/50]  eta: 0:00:46  loss: 3.1298 (3.2012)  acc1: 35.2000 (36.2286)  acc5: 61.6000 (60.1905)  time: 0.9685  data: 0.9486  max mem: 2905
Test:  [30/50]  eta: 0:00:27  loss: 3.3673 (3.2551)  acc1: 31.2000 (35.0194)  acc5: 56.0000 (59.2774)  time: 0.9817  data: 0.9614  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 3.5322 (3.3293)  acc1: 30.4000 (34.4000)  acc5: 53.6000 (58.2049)  time: 0.7766  data: 0.7562  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 3.2877 (3.3151)  acc1: 32.8000 (34.6560)  acc5: 58.4000 (58.5920)  time: 0.7002  data: 0.6775  max mem: 2905
Test: Total time: 0:00:50 (1.0199 s / it)
* Acc@1 34.898 Acc@5 59.410 loss 3.291
Accuracy of the model on the 50000 test images: 34.9%
Max accuracy: 50.54%
Epoch: [81]  [  0/625]  eta: 3:42:44  lr: 0.003550  min_lr: 0.003550  loss: 3.3862 (3.3862)  class_acc: 0.4492 (0.4492)  weight_decay: 0.0500 (0.0500)  time: 21.3832  data: 18.7355  max mem: 2905
Epoch: [81]  [200/625]  eta: 0:14:18  lr: 0.003545  min_lr: 0.003545  loss: 3.1863 (3.2257)  class_acc: 0.4844 (0.4772)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1731 (2.5033)  time: 2.0882  data: 0.0008  max mem: 2905
Epoch: [81]  [400/625]  eta: 0:07:24  lr: 0.003541  min_lr: 0.003541  loss: 3.2344 (3.2386)  class_acc: 0.4609 (0.4743)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4606 (2.4067)  time: 1.9807  data: 0.0108  max mem: 2905
Epoch: [81]  [600/625]  eta: 0:00:49  lr: 0.003536  min_lr: 0.003536  loss: 3.2765 (3.2428)  class_acc: 0.4688 (0.4739)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3574 (2.4149)  time: 2.0867  data: 0.0097  max mem: 2905
Epoch: [81]  [624/625]  eta: 0:00:01  lr: 0.003535  min_lr: 0.003535  loss: 3.2719 (3.2434)  class_acc: 0.4570 (0.4738)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0595 (2.3979)  time: 0.7564  data: 0.0017  max mem: 2905
Epoch: [81] Total time: 0:20:11 (1.9391 s / it)
Averaged stats: lr: 0.003535  min_lr: 0.003535  loss: 3.2719 (3.2360)  class_acc: 0.4570 (0.4745)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0595 (2.3979)
Test:  [ 0/50]  eta: 0:10:16  loss: 2.5239 (2.5239)  acc1: 42.4000 (42.4000)  acc5: 71.2000 (71.2000)  time: 12.3235  data: 12.2984  max mem: 2905
Test:  [10/50]  eta: 0:01:30  loss: 2.4535 (2.4161)  acc1: 45.6000 (46.6909)  acc5: 71.2000 (72.7273)  time: 2.2552  data: 2.2328  max mem: 2905
Test:  [20/50]  eta: 0:00:53  loss: 2.5067 (2.5621)  acc1: 45.6000 (44.9905)  acc5: 71.2000 (71.0476)  time: 1.2678  data: 1.2473  max mem: 2905
Test:  [30/50]  eta: 0:00:30  loss: 2.6581 (2.5426)  acc1: 43.2000 (45.3161)  acc5: 70.4000 (71.0452)  time: 1.1491  data: 1.1294  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 2.5609 (2.5721)  acc1: 44.0000 (45.0927)  acc5: 68.0000 (70.3610)  time: 0.7466  data: 0.7267  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.5889 (2.5729)  acc1: 48.0000 (45.2000)  acc5: 67.2000 (70.2080)  time: 0.6185  data: 0.5982  max mem: 2905
Test: Total time: 0:00:53 (1.0656 s / it)
* Acc@1 45.720 Acc@5 70.684 loss 2.557
Accuracy of the model on the 50000 test images: 45.7%
Max accuracy: 50.54%
Epoch: [82]  [  0/625]  eta: 3:50:20  lr: 0.003535  min_lr: 0.003535  loss: 3.2080 (3.2080)  class_acc: 0.4844 (0.4844)  weight_decay: 0.0500 (0.0500)  time: 22.1127  data: 18.8227  max mem: 2905
Epoch: [82]  [200/625]  eta: 0:14:37  lr: 0.003531  min_lr: 0.003531  loss: 3.2304 (3.2092)  class_acc: 0.4805 (0.4800)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3143 (2.3862)  time: 2.0445  data: 0.1168  max mem: 2905
Epoch: [82]  [400/625]  eta: 0:07:32  lr: 0.003526  min_lr: 0.003526  loss: 3.1940 (3.2177)  class_acc: 0.4844 (0.4777)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.6919 (2.2969)  time: 1.9838  data: 0.0009  max mem: 2905
Epoch: [82]  [600/625]  eta: 0:00:50  lr: 0.003521  min_lr: 0.003521  loss: 3.2946 (3.2266)  class_acc: 0.4609 (0.4757)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1730 (2.2894)  time: 2.0316  data: 0.0008  max mem: 2905
Epoch: [82]  [624/625]  eta: 0:00:01  lr: 0.003521  min_lr: 0.003521  loss: 3.2061 (3.2260)  class_acc: 0.4844 (0.4760)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7533 (2.3053)  time: 0.7690  data: 0.0012  max mem: 2905
Epoch: [82] Total time: 0:20:21 (1.9552 s / it)
Averaged stats: lr: 0.003521  min_lr: 0.003521  loss: 3.2061 (3.2329)  class_acc: 0.4844 (0.4750)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7533 (2.3053)
Test:  [ 0/50]  eta: 0:09:23  loss: 3.6125 (3.6125)  acc1: 28.0000 (28.0000)  acc5: 55.2000 (55.2000)  time: 11.2742  data: 11.2477  max mem: 2905
Test:  [10/50]  eta: 0:01:10  loss: 2.8790 (2.8529)  acc1: 41.6000 (41.3818)  acc5: 66.4000 (65.8182)  time: 1.7549  data: 1.7351  max mem: 2905
Test:  [20/50]  eta: 0:00:39  loss: 2.8978 (2.9828)  acc1: 38.4000 (38.8190)  acc5: 65.6000 (64.8381)  time: 0.8098  data: 0.7904  max mem: 2905
Test:  [30/50]  eta: 0:00:24  loss: 3.1137 (3.0085)  acc1: 36.8000 (38.6323)  acc5: 64.0000 (64.3097)  time: 0.9049  data: 0.8851  max mem: 2905
Test:  [40/50]  eta: 0:00:10  loss: 3.1587 (3.1080)  acc1: 35.2000 (37.4439)  acc5: 61.6000 (62.7902)  time: 0.7714  data: 0.7525  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 3.1427 (3.0892)  acc1: 35.2000 (37.3440)  acc5: 61.6000 (63.0720)  time: 0.4237  data: 0.4056  max mem: 2905
Test: Total time: 0:00:46 (0.9204 s / it)
* Acc@1 38.268 Acc@5 63.232 loss 3.059
Accuracy of the model on the 50000 test images: 38.3%
Max accuracy: 50.54%
Epoch: [83]  [  0/625]  eta: 4:56:25  lr: 0.003521  min_lr: 0.003521  loss: 3.0757 (3.0757)  class_acc: 0.5312 (0.5312)  weight_decay: 0.0500 (0.0500)  time: 28.4562  data: 16.8971  max mem: 2905
Epoch: [83]  [200/625]  eta: 0:14:12  lr: 0.003516  min_lr: 0.003516  loss: 3.1774 (3.2202)  class_acc: 0.4766 (0.4809)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1348 (2.3409)  time: 1.9362  data: 0.0006  max mem: 2905
Epoch: [83]  [400/625]  eta: 0:07:24  lr: 0.003512  min_lr: 0.003512  loss: 3.2603 (3.2223)  class_acc: 0.4688 (0.4788)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2746 (2.3557)  time: 1.9203  data: 0.0005  max mem: 2905
Epoch: [83]  [600/625]  eta: 0:00:49  lr: 0.003507  min_lr: 0.003507  loss: 3.2580 (3.2315)  class_acc: 0.4727 (0.4765)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0450 (2.4248)  time: 2.0852  data: 0.0006  max mem: 2905
Epoch: [83]  [624/625]  eta: 0:00:01  lr: 0.003506  min_lr: 0.003506  loss: 3.2510 (3.2318)  class_acc: 0.4609 (0.4762)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0070 (2.4234)  time: 1.0125  data: 0.0015  max mem: 2905
Epoch: [83] Total time: 0:20:08 (1.9342 s / it)
Averaged stats: lr: 0.003506  min_lr: 0.003506  loss: 3.2510 (3.2308)  class_acc: 0.4609 (0.4760)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0070 (2.4234)
Test:  [ 0/50]  eta: 0:09:37  loss: 3.5897 (3.5897)  acc1: 25.6000 (25.6000)  acc5: 56.8000 (56.8000)  time: 11.5551  data: 11.5257  max mem: 2905
Test:  [10/50]  eta: 0:01:25  loss: 2.8603 (2.8662)  acc1: 44.8000 (42.2545)  acc5: 67.2000 (66.5455)  time: 2.1307  data: 2.1110  max mem: 2905
Test:  [20/50]  eta: 0:00:51  loss: 2.9121 (2.9307)  acc1: 41.6000 (40.1905)  acc5: 67.2000 (65.7524)  time: 1.2357  data: 1.2163  max mem: 2905
Test:  [30/50]  eta: 0:00:30  loss: 2.9853 (2.9078)  acc1: 39.2000 (40.7484)  acc5: 64.8000 (65.8839)  time: 1.2107  data: 1.1904  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 2.9584 (2.9032)  acc1: 41.6000 (40.8000)  acc5: 64.8000 (66.1659)  time: 0.7657  data: 0.7460  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.9494 (2.9020)  acc1: 43.2000 (41.2960)  acc5: 68.0000 (66.2400)  time: 0.6063  data: 0.5869  max mem: 2905
Test: Total time: 0:00:52 (1.0557 s / it)
* Acc@1 41.544 Acc@5 66.228 loss 2.859
Accuracy of the model on the 50000 test images: 41.5%
Max accuracy: 50.54%
Epoch: [84]  [  0/625]  eta: 3:25:58  lr: 0.003506  min_lr: 0.003506  loss: 3.3097 (3.3097)  class_acc: 0.4180 (0.4180)  weight_decay: 0.0500 (0.0500)  time: 19.7730  data: 18.7967  max mem: 2905
Epoch: [84]  [200/625]  eta: 0:13:43  lr: 0.003502  min_lr: 0.003502  loss: 3.2494 (3.2083)  class_acc: 0.4844 (0.4841)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9331 (2.2869)  time: 1.8785  data: 0.4316  max mem: 2905
Epoch: [84]  [400/625]  eta: 0:07:17  lr: 0.003497  min_lr: 0.003497  loss: 3.2281 (3.2140)  class_acc: 0.4727 (0.4815)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0059 (2.2879)  time: 2.2076  data: 0.0110  max mem: 2905
Epoch: [84]  [600/625]  eta: 0:00:49  lr: 0.003492  min_lr: 0.003492  loss: 3.2498 (3.2230)  class_acc: 0.4688 (0.4793)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1001 (2.3016)  time: 1.9811  data: 0.0083  max mem: 2905
Epoch: [84]  [624/625]  eta: 0:00:01  lr: 0.003491  min_lr: 0.003491  loss: 3.2252 (3.2234)  class_acc: 0.4570 (0.4787)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2111 (2.3185)  time: 0.6303  data: 0.0013  max mem: 2905
Epoch: [84] Total time: 0:20:20 (1.9534 s / it)
Averaged stats: lr: 0.003491  min_lr: 0.003491  loss: 3.2252 (3.2286)  class_acc: 0.4570 (0.4766)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2111 (2.3185)
Test:  [ 0/50]  eta: 0:10:26  loss: 2.7145 (2.7145)  acc1: 40.0000 (40.0000)  acc5: 72.0000 (72.0000)  time: 12.5353  data: 12.4993  max mem: 2905
Test:  [10/50]  eta: 0:01:23  loss: 2.7145 (2.6205)  acc1: 44.0000 (45.0182)  acc5: 68.8000 (70.8364)  time: 2.0960  data: 2.0761  max mem: 2905
Test:  [20/50]  eta: 0:00:48  loss: 2.7654 (2.8310)  acc1: 42.4000 (41.6000)  acc5: 68.8000 (67.7714)  time: 1.0872  data: 1.0679  max mem: 2905
Test:  [30/50]  eta: 0:00:29  loss: 2.8514 (2.8097)  acc1: 39.2000 (42.2710)  acc5: 64.8000 (67.5355)  time: 1.0979  data: 1.0783  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 2.7876 (2.8573)  acc1: 41.6000 (41.6781)  acc5: 63.2000 (66.7902)  time: 0.9016  data: 0.8827  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.7876 (2.8668)  acc1: 41.6000 (41.3600)  acc5: 66.4000 (66.8000)  time: 0.8606  data: 0.8412  max mem: 2905
Test: Total time: 0:00:55 (1.1005 s / it)
* Acc@1 42.174 Acc@5 67.112 loss 2.809
Accuracy of the model on the 50000 test images: 42.2%
Max accuracy: 50.54%
Epoch: [85]  [  0/625]  eta: 3:52:01  lr: 0.003491  min_lr: 0.003491  loss: 3.1183 (3.1183)  class_acc: 0.5000 (0.5000)  weight_decay: 0.0500 (0.0500)  time: 22.2746  data: 15.9025  max mem: 2905
Epoch: [85]  [200/625]  eta: 0:14:20  lr: 0.003487  min_lr: 0.003487  loss: 3.1167 (3.1996)  class_acc: 0.4961 (0.4793)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3758 (2.5677)  time: 2.0041  data: 0.0007  max mem: 2905
Epoch: [85]  [400/625]  eta: 0:07:27  lr: 0.003482  min_lr: 0.003482  loss: 3.2080 (3.2102)  class_acc: 0.4688 (0.4795)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5124 (2.3747)  time: 2.0525  data: 0.0007  max mem: 2905
Epoch: [85]  [600/625]  eta: 0:00:49  lr: 0.003477  min_lr: 0.003477  loss: 3.1954 (3.2200)  class_acc: 0.4727 (0.4782)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2954 (2.4543)  time: 2.0844  data: 0.0007  max mem: 2905
Epoch: [85]  [624/625]  eta: 0:00:01  lr: 0.003476  min_lr: 0.003476  loss: 3.2326 (3.2217)  class_acc: 0.4688 (0.4780)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9679 (2.4388)  time: 0.7124  data: 0.0069  max mem: 2905
Epoch: [85] Total time: 0:20:33 (1.9735 s / it)
Averaged stats: lr: 0.003476  min_lr: 0.003476  loss: 3.2326 (3.2250)  class_acc: 0.4688 (0.4770)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9679 (2.4388)
Test:  [ 0/50]  eta: 0:11:07  loss: 2.4857 (2.4857)  acc1: 43.2000 (43.2000)  acc5: 72.0000 (72.0000)  time: 13.3551  data: 13.3195  max mem: 2905
Test:  [10/50]  eta: 0:01:29  loss: 2.6886 (2.6757)  acc1: 43.2000 (44.5091)  acc5: 68.8000 (69.3818)  time: 2.2306  data: 2.2072  max mem: 2905
Test:  [20/50]  eta: 0:00:51  loss: 2.8033 (2.8298)  acc1: 37.6000 (39.9238)  acc5: 67.2000 (67.3524)  time: 1.1288  data: 1.1077  max mem: 2905
Test:  [30/50]  eta: 0:00:28  loss: 2.9182 (2.8089)  acc1: 36.8000 (40.6452)  acc5: 65.6000 (66.9936)  time: 0.9553  data: 0.9353  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 2.9188 (2.8687)  acc1: 36.8000 (39.7463)  acc5: 64.0000 (66.4000)  time: 0.5317  data: 0.5123  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 2.8870 (2.8784)  acc1: 39.2000 (40.1920)  acc5: 64.8000 (66.0640)  time: 0.4663  data: 0.4472  max mem: 2905
Test: Total time: 0:00:47 (0.9431 s / it)
* Acc@1 41.010 Acc@5 66.718 loss 2.847
Accuracy of the model on the 50000 test images: 41.0%
Max accuracy: 50.54%
Epoch: [86]  [  0/625]  eta: 3:39:37  lr: 0.003476  min_lr: 0.003476  loss: 3.1238 (3.1238)  class_acc: 0.5234 (0.5234)  weight_decay: 0.0500 (0.0500)  time: 21.0835  data: 16.5965  max mem: 2905
Epoch: [86]  [200/625]  eta: 0:14:14  lr: 0.003472  min_lr: 0.003472  loss: 3.2137 (3.2178)  class_acc: 0.4727 (0.4777)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7342 (2.3453)  time: 1.8250  data: 0.0258  max mem: 2905
Epoch: [86]  [400/625]  eta: 0:07:19  lr: 0.003467  min_lr: 0.003467  loss: 3.2722 (3.2242)  class_acc: 0.4609 (0.4762)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1211 (2.3977)  time: 1.9427  data: 0.0073  max mem: 2905
Epoch: [86]  [600/625]  eta: 0:00:48  lr: 0.003462  min_lr: 0.003462  loss: 3.2211 (3.2284)  class_acc: 0.4844 (0.4764)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1045 (2.4542)  time: 1.9558  data: 0.0008  max mem: 2905
Epoch: [86]  [624/625]  eta: 0:00:01  lr: 0.003461  min_lr: 0.003461  loss: 3.2572 (3.2290)  class_acc: 0.4648 (0.4761)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0516 (2.4424)  time: 0.7033  data: 0.0027  max mem: 2905
Epoch: [86] Total time: 0:19:43 (1.8941 s / it)
Averaged stats: lr: 0.003461  min_lr: 0.003461  loss: 3.2572 (3.2208)  class_acc: 0.4648 (0.4785)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0516 (2.4424)
Test:  [ 0/50]  eta: 0:10:38  loss: 2.6977 (2.6977)  acc1: 35.2000 (35.2000)  acc5: 70.4000 (70.4000)  time: 12.7678  data: 12.7387  max mem: 2905
Test:  [10/50]  eta: 0:01:30  loss: 2.6977 (2.6980)  acc1: 44.0000 (43.5636)  acc5: 68.8000 (69.1636)  time: 2.2671  data: 2.2468  max mem: 2905
Test:  [20/50]  eta: 0:00:54  loss: 2.8232 (2.8710)  acc1: 41.6000 (41.3714)  acc5: 65.6000 (66.1714)  time: 1.2520  data: 1.2306  max mem: 2905
Test:  [30/50]  eta: 0:00:30  loss: 2.9138 (2.8949)  acc1: 39.2000 (41.1355)  acc5: 64.0000 (65.3936)  time: 1.1397  data: 1.1182  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 2.9052 (2.9008)  acc1: 40.0000 (41.3854)  acc5: 64.8000 (65.3268)  time: 0.6785  data: 0.6589  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.8334 (2.8894)  acc1: 41.6000 (41.5840)  acc5: 66.4000 (65.7120)  time: 0.6200  data: 0.6009  max mem: 2905
Test: Total time: 0:00:52 (1.0405 s / it)
* Acc@1 41.826 Acc@5 66.430 loss 2.853
Accuracy of the model on the 50000 test images: 41.8%
Max accuracy: 50.54%
Epoch: [87]  [  0/625]  eta: 3:36:40  lr: 0.003461  min_lr: 0.003461  loss: 3.0572 (3.0572)  class_acc: 0.5000 (0.5000)  weight_decay: 0.0500 (0.0500)  time: 20.8011  data: 20.5531  max mem: 2905
Epoch: [87]  [200/625]  eta: 0:13:55  lr: 0.003456  min_lr: 0.003456  loss: 3.1673 (3.1950)  class_acc: 0.4883 (0.4839)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8527 (inf)  time: 1.8162  data: 0.5992  max mem: 2905
Epoch: [87]  [400/625]  eta: 0:07:17  lr: 0.003451  min_lr: 0.003451  loss: 3.2221 (3.2059)  class_acc: 0.4727 (0.4820)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4221 (inf)  time: 1.8759  data: 0.0911  max mem: 2905
Epoch: [87]  [600/625]  eta: 0:00:49  lr: 0.003446  min_lr: 0.003446  loss: 3.2460 (3.2084)  class_acc: 0.4766 (0.4806)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8130 (inf)  time: 1.9575  data: 0.0009  max mem: 2905
Epoch: [87]  [624/625]  eta: 0:00:01  lr: 0.003446  min_lr: 0.003446  loss: 3.2013 (3.2090)  class_acc: 0.4727 (0.4803)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.5561 (inf)  time: 0.7937  data: 0.0018  max mem: 2905
Epoch: [87] Total time: 0:20:05 (1.9285 s / it)
Averaged stats: lr: 0.003446  min_lr: 0.003446  loss: 3.2013 (3.2166)  class_acc: 0.4727 (0.4791)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.5561 (inf)
Test:  [ 0/50]  eta: 0:09:48  loss: 2.4731 (2.4731)  acc1: 45.6000 (45.6000)  acc5: 71.2000 (71.2000)  time: 11.7673  data: 11.7303  max mem: 2905
Test:  [10/50]  eta: 0:01:25  loss: 2.5290 (2.6485)  acc1: 45.6000 (45.3091)  acc5: 69.6000 (69.6000)  time: 2.1425  data: 2.1228  max mem: 2905
Test:  [20/50]  eta: 0:00:52  loss: 2.8579 (2.7854)  acc1: 40.0000 (42.4000)  acc5: 66.4000 (67.6952)  time: 1.2531  data: 1.2341  max mem: 2905
Test:  [30/50]  eta: 0:00:31  loss: 2.9216 (2.8183)  acc1: 39.2000 (41.9355)  acc5: 65.6000 (66.8387)  time: 1.2380  data: 1.2187  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 2.9664 (2.8576)  acc1: 39.2000 (41.2683)  acc5: 64.8000 (66.1073)  time: 0.7718  data: 0.7529  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.8881 (2.8694)  acc1: 39.2000 (41.2800)  acc5: 65.6000 (66.2560)  time: 0.6690  data: 0.6495  max mem: 2905
Test: Total time: 0:00:53 (1.0618 s / it)
* Acc@1 41.934 Acc@5 66.992 loss 2.814
Accuracy of the model on the 50000 test images: 41.9%
Max accuracy: 50.54%
Epoch: [88]  [  0/625]  eta: 3:43:40  lr: 0.003446  min_lr: 0.003446  loss: 3.3014 (3.3014)  class_acc: 0.4727 (0.4727)  weight_decay: 0.0500 (0.0500)  time: 21.4722  data: 21.3492  max mem: 2905
Epoch: [88]  [200/625]  eta: 0:14:12  lr: 0.003441  min_lr: 0.003441  loss: 3.2173 (3.2094)  class_acc: 0.4609 (0.4784)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0129 (2.0891)  time: 1.9693  data: 0.0255  max mem: 2905
Epoch: [88]  [400/625]  eta: 0:07:26  lr: 0.003436  min_lr: 0.003436  loss: 3.1962 (3.2177)  class_acc: 0.4727 (0.4780)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4191 (2.2923)  time: 1.8834  data: 0.0013  max mem: 2905
Epoch: [88]  [600/625]  eta: 0:00:49  lr: 0.003431  min_lr: 0.003431  loss: 3.2292 (3.2176)  class_acc: 0.4648 (0.4785)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2967 (2.3386)  time: 1.9979  data: 0.0007  max mem: 2905
Epoch: [88]  [624/625]  eta: 0:00:01  lr: 0.003430  min_lr: 0.003430  loss: 3.1960 (3.2165)  class_acc: 0.4688 (0.4788)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.6856 (2.3120)  time: 0.7321  data: 0.0117  max mem: 2905
Epoch: [88] Total time: 0:20:32 (1.9727 s / it)
Averaged stats: lr: 0.003430  min_lr: 0.003430  loss: 3.1960 (3.2131)  class_acc: 0.4688 (0.4796)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.6856 (2.3120)
Test:  [ 0/50]  eta: 0:10:33  loss: 2.0753 (2.0753)  acc1: 56.0000 (56.0000)  acc5: 82.4000 (82.4000)  time: 12.6689  data: 12.6069  max mem: 2905
Test:  [10/50]  eta: 0:01:23  loss: 2.2820 (2.3151)  acc1: 54.4000 (51.2727)  acc5: 74.4000 (74.2545)  time: 2.0962  data: 2.0732  max mem: 2905
Test:  [20/50]  eta: 0:00:49  loss: 2.3739 (2.4700)  acc1: 48.0000 (47.5429)  acc5: 72.8000 (72.8762)  time: 1.0980  data: 1.0784  max mem: 2905
Test:  [30/50]  eta: 0:00:29  loss: 2.5386 (2.4788)  acc1: 44.0000 (47.4581)  acc5: 71.2000 (72.3613)  time: 1.1526  data: 1.1328  max mem: 2905
Test:  [40/50]  eta: 0:00:13  loss: 2.5386 (2.4966)  acc1: 44.0000 (47.2000)  acc5: 70.4000 (72.0781)  time: 0.9380  data: 0.9169  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.5597 (2.5066)  acc1: 45.6000 (47.2160)  acc5: 68.8000 (71.7920)  time: 0.8518  data: 0.8287  max mem: 2905
Test: Total time: 0:00:54 (1.0840 s / it)
* Acc@1 47.640 Acc@5 72.136 loss 2.453
Accuracy of the model on the 50000 test images: 47.6%
Max accuracy: 50.54%
Epoch: [89]  [  0/625]  eta: 3:27:03  lr: 0.003430  min_lr: 0.003430  loss: 3.2908 (3.2908)  class_acc: 0.4609 (0.4609)  weight_decay: 0.0500 (0.0500)  time: 19.8774  data: 19.7514  max mem: 2905
Epoch: [89]  [200/625]  eta: 0:14:34  lr: 0.003425  min_lr: 0.003425  loss: 3.2745 (3.2132)  class_acc: 0.4648 (0.4771)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5580 (2.5102)  time: 2.0330  data: 0.0057  max mem: 2905
Epoch: [89]  [400/625]  eta: 0:07:23  lr: 0.003420  min_lr: 0.003420  loss: 3.1864 (3.2044)  class_acc: 0.4766 (0.4797)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.6811 (2.4566)  time: 1.8411  data: 0.0225  max mem: 2905
Epoch: [89]  [600/625]  eta: 0:00:48  lr: 0.003415  min_lr: 0.003415  loss: 3.2201 (3.2147)  class_acc: 0.4844 (0.4782)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9170 (2.4484)  time: 1.9983  data: 0.0006  max mem: 2905
Epoch: [89]  [624/625]  eta: 0:00:01  lr: 0.003414  min_lr: 0.003414  loss: 3.2429 (3.2162)  class_acc: 0.4648 (0.4780)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4656 (2.4715)  time: 0.7636  data: 0.0013  max mem: 2905
Epoch: [89] Total time: 0:19:54 (1.9114 s / it)
Averaged stats: lr: 0.003414  min_lr: 0.003414  loss: 3.2429 (3.2108)  class_acc: 0.4648 (0.4797)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4656 (2.4715)
Test:  [ 0/50]  eta: 0:10:53  loss: 2.9856 (2.9856)  acc1: 36.0000 (36.0000)  acc5: 58.4000 (58.4000)  time: 13.0795  data: 13.0515  max mem: 2905
Test:  [10/50]  eta: 0:01:27  loss: 2.6802 (2.7531)  acc1: 41.6000 (42.2545)  acc5: 69.6000 (67.0545)  time: 2.1886  data: 2.1679  max mem: 2905
Test:  [20/50]  eta: 0:00:51  loss: 2.7576 (2.8603)  acc1: 40.0000 (41.1429)  acc5: 66.4000 (66.4762)  time: 1.1448  data: 1.1254  max mem: 2905
Test:  [30/50]  eta: 0:00:26  loss: 2.8800 (2.8196)  acc1: 40.0000 (41.7548)  acc5: 66.4000 (66.9161)  time: 0.8852  data: 0.8659  max mem: 2905
Test:  [40/50]  eta: 0:00:10  loss: 2.7411 (2.8420)  acc1: 39.2000 (41.1317)  acc5: 66.4000 (66.5561)  time: 0.4482  data: 0.4290  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 2.8411 (2.8381)  acc1: 39.2000 (41.1040)  acc5: 67.2000 (66.6400)  time: 0.4669  data: 0.4474  max mem: 2905
Test: Total time: 0:00:46 (0.9233 s / it)
* Acc@1 42.196 Acc@5 67.034 loss 2.805
Accuracy of the model on the 50000 test images: 42.2%
Max accuracy: 50.54%
Epoch: [90]  [  0/625]  eta: 3:24:46  lr: 0.003414  min_lr: 0.003414  loss: 3.3031 (3.3031)  class_acc: 0.4492 (0.4492)  weight_decay: 0.0500 (0.0500)  time: 19.6577  data: 18.9060  max mem: 2905
Epoch: [90]  [200/625]  eta: 0:14:05  lr: 0.003409  min_lr: 0.003409  loss: 3.1990 (3.1914)  class_acc: 0.4727 (0.4836)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1038 (2.3971)  time: 1.8025  data: 0.0006  max mem: 2905
Epoch: [90]  [400/625]  eta: 0:07:25  lr: 0.003404  min_lr: 0.003404  loss: 3.1987 (3.1960)  class_acc: 0.4883 (0.4836)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0117 (2.3809)  time: 2.0952  data: 0.0016  max mem: 2905
Epoch: [90]  [600/625]  eta: 0:00:49  lr: 0.003399  min_lr: 0.003399  loss: 3.2620 (3.2044)  class_acc: 0.4609 (0.4809)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0815 (2.4661)  time: 2.0855  data: 0.0242  max mem: 2905
Epoch: [90]  [624/625]  eta: 0:00:01  lr: 0.003398  min_lr: 0.003398  loss: 3.2434 (3.2058)  class_acc: 0.4688 (0.4806)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5561 (2.4694)  time: 0.9347  data: 0.0013  max mem: 2905
Epoch: [90] Total time: 0:20:16 (1.9462 s / it)
Averaged stats: lr: 0.003398  min_lr: 0.003398  loss: 3.2434 (3.2069)  class_acc: 0.4688 (0.4805)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5561 (2.4694)
Test:  [ 0/50]  eta: 0:10:37  loss: 2.5887 (2.5887)  acc1: 43.2000 (43.2000)  acc5: 74.4000 (74.4000)  time: 12.7457  data: 12.7187  max mem: 2905
Test:  [10/50]  eta: 0:01:19  loss: 2.5916 (2.7249)  acc1: 44.0000 (44.7273)  acc5: 70.4000 (68.5091)  time: 1.9900  data: 1.9712  max mem: 2905
Test:  [20/50]  eta: 0:00:46  loss: 2.9536 (2.8927)  acc1: 38.4000 (40.6476)  acc5: 65.6000 (66.2857)  time: 0.9816  data: 0.9635  max mem: 2905
Test:  [30/50]  eta: 0:00:27  loss: 2.9536 (2.8755)  acc1: 37.6000 (40.1290)  acc5: 64.8000 (66.0129)  time: 1.0233  data: 1.0045  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 2.8274 (2.8910)  acc1: 38.4000 (40.0585)  acc5: 64.8000 (65.6781)  time: 0.9459  data: 0.9260  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.8944 (2.8773)  acc1: 39.2000 (40.0160)  acc5: 66.4000 (65.8240)  time: 0.9833  data: 0.9642  max mem: 2905
Test: Total time: 0:00:57 (1.1524 s / it)
* Acc@1 40.932 Acc@5 66.728 loss 2.844
Accuracy of the model on the 50000 test images: 40.9%
Max accuracy: 50.54%
Epoch: [91]  [  0/625]  eta: 3:44:45  lr: 0.003398  min_lr: 0.003398  loss: 3.2703 (3.2703)  class_acc: 0.4453 (0.4453)  weight_decay: 0.0500 (0.0500)  time: 21.5764  data: 19.1172  max mem: 2905
Epoch: [91]  [200/625]  eta: 0:15:07  lr: 0.003393  min_lr: 0.003393  loss: 3.1338 (3.1884)  class_acc: 0.4805 (0.4838)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8080 (2.3157)  time: 2.0237  data: 0.0512  max mem: 2905
Epoch: [91]  [400/625]  eta: 0:07:39  lr: 0.003388  min_lr: 0.003388  loss: 3.1891 (3.1953)  class_acc: 0.4883 (0.4832)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.6608 (2.4206)  time: 1.9308  data: 0.0008  max mem: 2905
Epoch: [91]  [600/625]  eta: 0:00:50  lr: 0.003383  min_lr: 0.003383  loss: 3.2319 (3.2035)  class_acc: 0.4766 (0.4809)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2538 (2.4311)  time: 1.9758  data: 0.0007  max mem: 2905
Epoch: [91]  [624/625]  eta: 0:00:02  lr: 0.003382  min_lr: 0.003382  loss: 3.2151 (3.2043)  class_acc: 0.4766 (0.4808)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7773 (2.4272)  time: 0.3949  data: 0.0030  max mem: 2905
Epoch: [91] Total time: 0:20:53 (2.0049 s / it)
Averaged stats: lr: 0.003382  min_lr: 0.003382  loss: 3.2151 (3.2037)  class_acc: 0.4766 (0.4814)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7773 (2.4272)
Test:  [ 0/50]  eta: 0:10:32  loss: 3.0355 (3.0355)  acc1: 36.8000 (36.8000)  acc5: 64.8000 (64.8000)  time: 12.6457  data: 12.6177  max mem: 2905
Test:  [10/50]  eta: 0:01:23  loss: 2.8191 (2.7846)  acc1: 40.8000 (42.3273)  acc5: 65.6000 (68.2182)  time: 2.0807  data: 2.0610  max mem: 2905
Test:  [20/50]  eta: 0:00:49  loss: 2.9319 (2.9184)  acc1: 38.4000 (40.6095)  acc5: 65.6000 (66.5905)  time: 1.0990  data: 1.0793  max mem: 2905
Test:  [30/50]  eta: 0:00:29  loss: 3.0517 (2.9091)  acc1: 39.2000 (40.2839)  acc5: 65.6000 (66.3484)  time: 1.1542  data: 1.1349  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 2.9775 (2.9410)  acc1: 39.2000 (40.1561)  acc5: 65.6000 (65.8341)  time: 0.9119  data: 0.8935  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.9188 (2.9157)  acc1: 39.2000 (40.4800)  acc5: 64.8000 (66.1440)  time: 0.8133  data: 0.7940  max mem: 2905
Test: Total time: 0:00:53 (1.0732 s / it)
* Acc@1 41.224 Acc@5 66.532 loss 2.869
Accuracy of the model on the 50000 test images: 41.2%
Max accuracy: 50.54%
Epoch: [92]  [  0/625]  eta: 4:09:38  lr: 0.003382  min_lr: 0.003382  loss: 3.1955 (3.1955)  class_acc: 0.5000 (0.5000)  weight_decay: 0.0500 (0.0500)  time: 23.9660  data: 19.1441  max mem: 2905
Epoch: [92]  [200/625]  eta: 0:15:13  lr: 0.003377  min_lr: 0.003377  loss: 3.0893 (3.1763)  class_acc: 0.4805 (0.4867)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2637 (2.2168)  time: 1.9879  data: 0.0055  max mem: 2905
Epoch: [92]  [400/625]  eta: 0:07:49  lr: 0.003372  min_lr: 0.003372  loss: 3.2306 (3.1971)  class_acc: 0.4766 (0.4825)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6521 (2.4301)  time: 2.1071  data: 0.0009  max mem: 2905
Epoch: [92]  [600/625]  eta: 0:00:51  lr: 0.003367  min_lr: 0.003367  loss: 3.1990 (3.2010)  class_acc: 0.4766 (0.4817)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4445 (2.4852)  time: 1.9103  data: 0.0009  max mem: 2905
Epoch: [92]  [624/625]  eta: 0:00:02  lr: 0.003366  min_lr: 0.003366  loss: 3.2054 (3.2019)  class_acc: 0.4727 (0.4812)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7489 (2.4519)  time: 0.5328  data: 0.0013  max mem: 2905
Epoch: [92] Total time: 0:20:51 (2.0020 s / it)
Averaged stats: lr: 0.003366  min_lr: 0.003366  loss: 3.2054 (3.2000)  class_acc: 0.4727 (0.4825)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7489 (2.4519)
Test:  [ 0/50]  eta: 0:09:50  loss: 2.7334 (2.7334)  acc1: 44.8000 (44.8000)  acc5: 71.2000 (71.2000)  time: 11.8003  data: 11.7737  max mem: 2905
Test:  [10/50]  eta: 0:01:12  loss: 2.5515 (2.4975)  acc1: 48.0000 (47.2000)  acc5: 69.6000 (70.4000)  time: 1.8221  data: 1.8021  max mem: 2905
Test:  [20/50]  eta: 0:00:41  loss: 2.6460 (2.6877)  acc1: 44.8000 (44.0000)  acc5: 69.6000 (68.6857)  time: 0.8466  data: 0.8265  max mem: 2905
Test:  [30/50]  eta: 0:00:24  loss: 2.6825 (2.6899)  acc1: 39.2000 (43.5871)  acc5: 68.0000 (68.8258)  time: 0.9090  data: 0.8884  max mem: 2905
Test:  [40/50]  eta: 0:00:10  loss: 2.7232 (2.7105)  acc1: 43.2000 (43.6293)  acc5: 68.0000 (68.6439)  time: 0.7509  data: 0.7314  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 2.8209 (2.7290)  acc1: 43.2000 (43.6000)  acc5: 67.2000 (68.3200)  time: 0.4287  data: 0.4102  max mem: 2905
Test: Total time: 0:00:46 (0.9390 s / it)
* Acc@1 44.554 Acc@5 69.216 loss 2.669
Accuracy of the model on the 50000 test images: 44.6%
Max accuracy: 50.54%
Epoch: [93]  [  0/625]  eta: 3:57:51  lr: 0.003366  min_lr: 0.003366  loss: 3.1277 (3.1277)  class_acc: 0.4844 (0.4844)  weight_decay: 0.0500 (0.0500)  time: 22.8344  data: 17.9874  max mem: 2905
Epoch: [93]  [200/625]  eta: 0:14:04  lr: 0.003361  min_lr: 0.003361  loss: 3.2206 (3.1788)  class_acc: 0.4844 (0.4901)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1617 (2.3268)  time: 1.9248  data: 0.0007  max mem: 2905
Epoch: [93]  [400/625]  eta: 0:07:21  lr: 0.003355  min_lr: 0.003355  loss: 3.1919 (3.1772)  class_acc: 0.4883 (0.4884)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3519 (2.3898)  time: 1.9549  data: 0.0008  max mem: 2905
Epoch: [93]  [600/625]  eta: 0:00:49  lr: 0.003350  min_lr: 0.003350  loss: 3.2102 (3.1923)  class_acc: 0.4727 (0.4846)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9046 (inf)  time: 2.0385  data: 0.0025  max mem: 2905
Epoch: [93]  [624/625]  eta: 0:00:01  lr: 0.003350  min_lr: 0.003350  loss: 3.1707 (3.1921)  class_acc: 0.4844 (0.4845)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7101 (inf)  time: 0.5544  data: 0.0012  max mem: 2905
Epoch: [93] Total time: 0:20:20 (1.9530 s / it)
Averaged stats: lr: 0.003350  min_lr: 0.003350  loss: 3.1707 (3.1948)  class_acc: 0.4844 (0.4839)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7101 (inf)
Test:  [ 0/50]  eta: 0:10:15  loss: 2.7892 (2.7892)  acc1: 38.4000 (38.4000)  acc5: 67.2000 (67.2000)  time: 12.3007  data: 12.2757  max mem: 2905
Test:  [10/50]  eta: 0:01:22  loss: 2.7373 (2.7327)  acc1: 45.6000 (43.3455)  acc5: 68.0000 (69.0182)  time: 2.0597  data: 2.0408  max mem: 2905
Test:  [20/50]  eta: 0:00:49  loss: 2.8712 (2.8986)  acc1: 40.0000 (40.1905)  acc5: 67.2000 (65.9429)  time: 1.1081  data: 1.0895  max mem: 2905
Test:  [30/50]  eta: 0:00:29  loss: 2.8939 (2.8708)  acc1: 39.2000 (41.1097)  acc5: 62.4000 (65.9355)  time: 1.1571  data: 1.1383  max mem: 2905
Test:  [40/50]  eta: 0:00:13  loss: 2.8631 (2.9068)  acc1: 41.6000 (40.8585)  acc5: 64.8000 (65.6781)  time: 1.0247  data: 1.0033  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.8560 (2.8971)  acc1: 40.0000 (40.7360)  acc5: 65.6000 (65.9200)  time: 0.9014  data: 0.8793  max mem: 2905
Test: Total time: 0:00:56 (1.1287 s / it)
* Acc@1 41.162 Acc@5 65.946 loss 2.865
Accuracy of the model on the 50000 test images: 41.2%
Max accuracy: 50.54%
Epoch: [94]  [  0/625]  eta: 3:41:05  lr: 0.003350  min_lr: 0.003350  loss: 3.2145 (3.2145)  class_acc: 0.4766 (0.4766)  weight_decay: 0.0500 (0.0500)  time: 21.2249  data: 19.2294  max mem: 2905
Epoch: [94]  [200/625]  eta: 0:14:09  lr: 0.003344  min_lr: 0.003344  loss: 3.1320 (3.1687)  class_acc: 0.4844 (0.4890)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8061 (2.2527)  time: 1.9739  data: 0.0007  max mem: 2905
Epoch: [94]  [400/625]  eta: 0:07:26  lr: 0.003339  min_lr: 0.003339  loss: 3.1528 (3.1754)  class_acc: 0.5000 (0.4877)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7392 (2.3678)  time: 1.9544  data: 0.0007  max mem: 2905
Epoch: [94]  [600/625]  eta: 0:00:49  lr: 0.003334  min_lr: 0.003334  loss: 3.2093 (3.1897)  class_acc: 0.4727 (0.4851)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7204 (2.2605)  time: 2.0430  data: 0.0007  max mem: 2905
Epoch: [94]  [624/625]  eta: 0:00:01  lr: 0.003333  min_lr: 0.003333  loss: 3.1833 (3.1901)  class_acc: 0.4805 (0.4847)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1688 (2.2645)  time: 0.9834  data: 0.0015  max mem: 2905
Epoch: [94] Total time: 0:20:10 (1.9365 s / it)
Averaged stats: lr: 0.003333  min_lr: 0.003333  loss: 3.1833 (3.1929)  class_acc: 0.4805 (0.4840)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1688 (2.2645)
Test:  [ 0/50]  eta: 0:11:08  loss: 3.7219 (3.7219)  acc1: 29.6000 (29.6000)  acc5: 52.8000 (52.8000)  time: 13.3745  data: 13.3477  max mem: 2905
Test:  [10/50]  eta: 0:01:20  loss: 3.7219 (3.7422)  acc1: 28.0000 (29.0182)  acc5: 52.8000 (53.0909)  time: 2.0177  data: 1.9984  max mem: 2905
Test:  [20/50]  eta: 0:00:44  loss: 3.8554 (3.9058)  acc1: 26.4000 (27.3905)  acc5: 48.8000 (50.5143)  time: 0.8972  data: 0.8786  max mem: 2905
Test:  [30/50]  eta: 0:00:25  loss: 4.0046 (3.9151)  acc1: 22.4000 (26.6839)  acc5: 46.4000 (50.2710)  time: 0.8295  data: 0.8103  max mem: 2905
Test:  [40/50]  eta: 0:00:10  loss: 4.0415 (3.9787)  acc1: 23.2000 (26.1659)  acc5: 47.2000 (49.7951)  time: 0.5859  data: 0.5670  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 4.1366 (4.0060)  acc1: 25.6000 (25.9360)  acc5: 48.0000 (49.4720)  time: 0.5389  data: 0.5200  max mem: 2905
Test: Total time: 0:00:43 (0.8734 s / it)
* Acc@1 26.544 Acc@5 50.080 loss 3.954
Accuracy of the model on the 50000 test images: 26.5%
Max accuracy: 50.54%
Epoch: [95]  [  0/625]  eta: 3:26:53  lr: 0.003333  min_lr: 0.003333  loss: 3.0119 (3.0119)  class_acc: 0.4961 (0.4961)  weight_decay: 0.0500 (0.0500)  time: 19.8616  data: 19.6932  max mem: 2905
Epoch: [95]  [200/625]  eta: 0:14:18  lr: 0.003327  min_lr: 0.003327  loss: 3.1781 (3.1674)  class_acc: 0.5000 (0.4885)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.5392 (2.3899)  time: 1.8658  data: 0.0010  max mem: 2905
Epoch: [95]  [400/625]  eta: 0:07:19  lr: 0.003322  min_lr: 0.003322  loss: 3.2103 (3.1855)  class_acc: 0.4844 (0.4858)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5359 (2.4348)  time: 1.9323  data: 0.0008  max mem: 2905
Epoch: [95]  [600/625]  eta: 0:00:48  lr: 0.003317  min_lr: 0.003317  loss: 3.1700 (3.1942)  class_acc: 0.4844 (0.4839)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6757 (2.4888)  time: 2.0826  data: 0.0356  max mem: 2905
Epoch: [95]  [624/625]  eta: 0:00:01  lr: 0.003316  min_lr: 0.003316  loss: 3.2174 (3.1947)  class_acc: 0.4688 (0.4837)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6382 (2.4824)  time: 0.8019  data: 0.0015  max mem: 2905
Epoch: [95] Total time: 0:19:56 (1.9137 s / it)
Averaged stats: lr: 0.003316  min_lr: 0.003316  loss: 3.2174 (3.1875)  class_acc: 0.4688 (0.4846)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6382 (2.4824)
Test:  [ 0/50]  eta: 0:10:36  loss: 3.2388 (3.2388)  acc1: 31.2000 (31.2000)  acc5: 57.6000 (57.6000)  time: 12.7226  data: 12.6828  max mem: 2905
Test:  [10/50]  eta: 0:01:25  loss: 2.7290 (2.8571)  acc1: 40.0000 (40.3636)  acc5: 66.4000 (67.0545)  time: 2.1277  data: 2.1026  max mem: 2905
Test:  [20/50]  eta: 0:00:49  loss: 3.0270 (2.9852)  acc1: 39.2000 (38.2857)  acc5: 64.0000 (64.6095)  time: 1.0855  data: 1.0627  max mem: 2905
Test:  [30/50]  eta: 0:00:28  loss: 3.0936 (3.0343)  acc1: 33.6000 (37.3161)  acc5: 62.4000 (63.3806)  time: 1.0356  data: 1.0144  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 3.2047 (3.0753)  acc1: 34.4000 (37.3463)  acc5: 60.0000 (62.5756)  time: 0.7524  data: 0.7319  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 3.1869 (3.0730)  acc1: 35.2000 (37.2160)  acc5: 61.6000 (62.6400)  time: 0.6658  data: 0.6445  max mem: 2905
Test: Total time: 0:00:50 (1.0064 s / it)
* Acc@1 38.344 Acc@5 63.160 loss 3.042
Accuracy of the model on the 50000 test images: 38.3%
Max accuracy: 50.54%
Epoch: [96]  [  0/625]  eta: 4:32:47  lr: 0.003316  min_lr: 0.003316  loss: 3.3384 (3.3384)  class_acc: 0.4531 (0.4531)  weight_decay: 0.0500 (0.0500)  time: 26.1884  data: 17.6749  max mem: 2905
Epoch: [96]  [200/625]  eta: 0:14:28  lr: 0.003311  min_lr: 0.003311  loss: 3.1485 (3.1673)  class_acc: 0.4922 (0.4882)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9368 (2.4397)  time: 1.9324  data: 0.0251  max mem: 2905
Epoch: [96]  [400/625]  eta: 0:07:27  lr: 0.003305  min_lr: 0.003305  loss: 3.1798 (3.1821)  class_acc: 0.4766 (0.4848)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4185 (2.3982)  time: 1.9602  data: 0.0455  max mem: 2905
Epoch: [96]  [600/625]  eta: 0:00:49  lr: 0.003300  min_lr: 0.003300  loss: 3.2417 (3.1865)  class_acc: 0.4727 (0.4850)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8276 (2.3432)  time: 1.9510  data: 0.0300  max mem: 2905
Epoch: [96]  [624/625]  eta: 0:00:01  lr: 0.003299  min_lr: 0.003299  loss: 3.2191 (3.1878)  class_acc: 0.4805 (0.4846)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1578 (2.3450)  time: 1.0348  data: 0.0176  max mem: 2905
Epoch: [96] Total time: 0:20:10 (1.9368 s / it)
Averaged stats: lr: 0.003299  min_lr: 0.003299  loss: 3.2191 (3.1868)  class_acc: 0.4805 (0.4853)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1578 (2.3450)
Test:  [ 0/50]  eta: 0:10:37  loss: 3.1733 (3.1733)  acc1: 42.4000 (42.4000)  acc5: 61.6000 (61.6000)  time: 12.7541  data: 12.7302  max mem: 2905
Test:  [10/50]  eta: 0:01:23  loss: 2.8607 (2.8430)  acc1: 40.8000 (42.1818)  acc5: 64.8000 (66.5455)  time: 2.0897  data: 2.0693  max mem: 2905
Test:  [20/50]  eta: 0:00:48  loss: 3.0559 (3.0544)  acc1: 37.6000 (38.6286)  acc5: 63.2000 (64.0762)  time: 1.0608  data: 1.0407  max mem: 2905
Test:  [30/50]  eta: 0:00:28  loss: 3.1425 (3.0267)  acc1: 37.6000 (39.2774)  acc5: 63.2000 (64.4645)  time: 1.0541  data: 1.0342  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 3.0507 (3.0390)  acc1: 39.2000 (38.9854)  acc5: 64.0000 (64.2537)  time: 0.7882  data: 0.7690  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 3.1531 (3.0842)  acc1: 36.8000 (38.2240)  acc5: 62.4000 (63.5360)  time: 0.6698  data: 0.6511  max mem: 2905
Test: Total time: 0:00:50 (1.0071 s / it)
* Acc@1 38.232 Acc@5 63.632 loss 3.045
Accuracy of the model on the 50000 test images: 38.2%
Max accuracy: 50.54%
Epoch: [97]  [  0/625]  eta: 3:34:01  lr: 0.003299  min_lr: 0.003299  loss: 3.0913 (3.0913)  class_acc: 0.5000 (0.5000)  weight_decay: 0.0500 (0.0500)  time: 20.5468  data: 17.2245  max mem: 2905
Epoch: [97]  [200/625]  eta: 0:14:08  lr: 0.003294  min_lr: 0.003294  loss: 3.2464 (3.1633)  class_acc: 0.4844 (0.4878)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.2161 (2.4480)  time: 1.8910  data: 0.0011  max mem: 2905
Epoch: [97]  [400/625]  eta: 0:07:19  lr: 0.003288  min_lr: 0.003288  loss: 3.1565 (3.1726)  class_acc: 0.4766 (0.4880)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9234 (2.3524)  time: 1.9757  data: 0.0007  max mem: 2905
Epoch: [97]  [600/625]  eta: 0:00:48  lr: 0.003283  min_lr: 0.003283  loss: 3.2316 (3.1810)  class_acc: 0.4805 (0.4878)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0122 (2.3987)  time: 1.9237  data: 0.0009  max mem: 2905
Epoch: [97]  [624/625]  eta: 0:00:01  lr: 0.003282  min_lr: 0.003282  loss: 3.1572 (3.1808)  class_acc: 0.4805 (0.4876)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0122 (2.4072)  time: 0.7896  data: 0.0016  max mem: 2905
Epoch: [97] Total time: 0:19:49 (1.9040 s / it)
Averaged stats: lr: 0.003282  min_lr: 0.003282  loss: 3.1572 (3.1837)  class_acc: 0.4805 (0.4861)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0122 (2.4072)
Test:  [ 0/50]  eta: 0:10:40  loss: 2.8172 (2.8172)  acc1: 43.2000 (43.2000)  acc5: 69.6000 (69.6000)  time: 12.8192  data: 12.7952  max mem: 2905
Test:  [10/50]  eta: 0:01:29  loss: 3.1881 (3.0690)  acc1: 36.8000 (38.4727)  acc5: 60.8000 (63.8545)  time: 2.2419  data: 2.2226  max mem: 2905
Test:  [20/50]  eta: 0:00:52  loss: 3.2060 (3.1854)  acc1: 34.4000 (36.0762)  acc5: 60.0000 (61.1810)  time: 1.2130  data: 1.1933  max mem: 2905
Test:  [30/50]  eta: 0:00:30  loss: 3.1958 (3.1799)  acc1: 33.6000 (36.3355)  acc5: 59.2000 (60.8000)  time: 1.0999  data: 1.0799  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 3.1758 (3.2466)  acc1: 34.4000 (35.2585)  acc5: 58.4000 (59.4341)  time: 0.6607  data: 0.6421  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 3.1973 (3.2212)  acc1: 34.4000 (35.4240)  acc5: 58.4000 (59.9680)  time: 0.5383  data: 0.5172  max mem: 2905
Test: Total time: 0:00:51 (1.0219 s / it)
* Acc@1 35.868 Acc@5 60.620 loss 3.195
Accuracy of the model on the 50000 test images: 35.9%
Max accuracy: 50.54%
Epoch: [98]  [  0/625]  eta: 3:45:23  lr: 0.003282  min_lr: 0.003282  loss: 2.9842 (2.9842)  class_acc: 0.5156 (0.5156)  weight_decay: 0.0500 (0.0500)  time: 21.6375  data: 18.5556  max mem: 2905
Epoch: [98]  [200/625]  eta: 0:14:19  lr: 0.003276  min_lr: 0.003276  loss: 3.1381 (3.1743)  class_acc: 0.4883 (0.4863)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8925 (2.2741)  time: 1.8770  data: 0.0333  max mem: 2905
Epoch: [98]  [400/625]  eta: 0:07:19  lr: 0.003271  min_lr: 0.003271  loss: 3.2362 (3.1812)  class_acc: 0.4805 (0.4859)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9335 (2.2470)  time: 1.9194  data: 0.0756  max mem: 2905
Epoch: [98]  [600/625]  eta: 0:00:48  lr: 0.003265  min_lr: 0.003265  loss: 3.1855 (3.1871)  class_acc: 0.4844 (0.4852)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2260 (2.3422)  time: 1.9480  data: 0.0007  max mem: 2905
Epoch: [98]  [624/625]  eta: 0:00:01  lr: 0.003265  min_lr: 0.003265  loss: 3.1908 (3.1877)  class_acc: 0.4805 (0.4851)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8568 (2.3188)  time: 0.9145  data: 0.0015  max mem: 2905
Epoch: [98] Total time: 0:19:52 (1.9085 s / it)
Averaged stats: lr: 0.003265  min_lr: 0.003265  loss: 3.1908 (3.1826)  class_acc: 0.4805 (0.4861)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8568 (2.3188)
Test:  [ 0/50]  eta: 0:09:33  loss: 3.4754 (3.4754)  acc1: 30.4000 (30.4000)  acc5: 54.4000 (54.4000)  time: 11.4651  data: 11.4382  max mem: 2905
Test:  [10/50]  eta: 0:01:17  loss: 2.8267 (2.9547)  acc1: 40.8000 (40.2909)  acc5: 62.4000 (62.6182)  time: 1.9294  data: 1.9096  max mem: 2905
Test:  [20/50]  eta: 0:00:45  loss: 3.1309 (3.1653)  acc1: 36.0000 (36.3429)  acc5: 62.4000 (61.0286)  time: 1.0109  data: 0.9920  max mem: 2905
Test:  [30/50]  eta: 0:00:26  loss: 3.3449 (3.2374)  acc1: 32.0000 (35.6129)  acc5: 57.6000 (60.0258)  time: 0.9970  data: 0.9780  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 3.3578 (3.2939)  acc1: 34.4000 (35.5707)  acc5: 56.8000 (59.3756)  time: 0.9319  data: 0.9104  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 3.2328 (3.2922)  acc1: 36.8000 (35.7440)  acc5: 58.4000 (59.3280)  time: 0.7029  data: 0.6817  max mem: 2905
Test: Total time: 0:00:54 (1.0803 s / it)
* Acc@1 36.370 Acc@5 60.698 loss 3.247
Accuracy of the model on the 50000 test images: 36.4%
Max accuracy: 50.54%
Epoch: [99]  [  0/625]  eta: 3:36:58  lr: 0.003265  min_lr: 0.003265  loss: 3.0258 (3.0258)  class_acc: 0.5117 (0.5117)  weight_decay: 0.0500 (0.0500)  time: 20.8301  data: 19.7825  max mem: 2905
Epoch: [99]  [200/625]  eta: 0:14:16  lr: 0.003259  min_lr: 0.003259  loss: 3.1685 (3.1541)  class_acc: 0.4883 (0.4904)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8130 (2.3655)  time: 1.7334  data: 0.0006  max mem: 2905
Epoch: [99]  [400/625]  eta: 0:07:23  lr: 0.003253  min_lr: 0.003253  loss: 3.2450 (3.1692)  class_acc: 0.4805 (0.4886)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3855 (2.4884)  time: 1.8393  data: 0.0006  max mem: 2905
Epoch: [99]  [600/625]  eta: 0:00:48  lr: 0.003248  min_lr: 0.003248  loss: 3.2114 (3.1752)  class_acc: 0.4766 (0.4871)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7437 (2.4589)  time: 1.9574  data: 0.0008  max mem: 2905
Epoch: [99]  [624/625]  eta: 0:00:01  lr: 0.003247  min_lr: 0.003247  loss: 3.2502 (3.1780)  class_acc: 0.4805 (0.4867)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.5675 (2.4343)  time: 0.4238  data: 0.0022  max mem: 2905
Epoch: [99] Total time: 0:19:53 (1.9098 s / it)
Averaged stats: lr: 0.003247  min_lr: 0.003247  loss: 3.2502 (3.1779)  class_acc: 0.4805 (0.4874)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.5675 (2.4343)
Test:  [ 0/50]  eta: 0:10:29  loss: 2.9894 (2.9894)  acc1: 42.4000 (42.4000)  acc5: 64.8000 (64.8000)  time: 12.5814  data: 12.5505  max mem: 2905
Test:  [10/50]  eta: 0:01:12  loss: 3.0311 (3.1152)  acc1: 37.6000 (38.6182)  acc5: 61.6000 (61.0182)  time: 1.8156  data: 1.7956  max mem: 2905
Test:  [20/50]  eta: 0:00:41  loss: 3.0638 (3.2249)  acc1: 37.6000 (36.8381)  acc5: 60.8000 (60.4191)  time: 0.8368  data: 0.8170  max mem: 2905
Test:  [30/50]  eta: 0:00:24  loss: 3.1582 (3.1819)  acc1: 36.0000 (37.0065)  acc5: 61.6000 (60.8774)  time: 0.9198  data: 0.9005  max mem: 2905
Test:  [40/50]  eta: 0:00:10  loss: 3.1503 (3.2181)  acc1: 36.0000 (36.3512)  acc5: 60.8000 (60.4098)  time: 0.7385  data: 0.7193  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 3.2797 (3.2487)  acc1: 34.4000 (35.9200)  acc5: 57.6000 (59.8080)  time: 0.4536  data: 0.4338  max mem: 2905
Test: Total time: 0:00:47 (0.9532 s / it)
* Acc@1 36.152 Acc@5 60.362 loss 3.227
Accuracy of the model on the 50000 test images: 36.2%
Max accuracy: 50.54%
Epoch: [100]  [  0/625]  eta: 3:39:06  lr: 0.003247  min_lr: 0.003247  loss: 3.1840 (3.1840)  class_acc: 0.4805 (0.4805)  weight_decay: 0.0500 (0.0500)  time: 21.0347  data: 19.0746  max mem: 2905
Epoch: [100]  [200/625]  eta: 0:14:15  lr: 0.003242  min_lr: 0.003242  loss: 3.1730 (3.1509)  class_acc: 0.4922 (0.4960)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9238 (inf)  time: 1.9224  data: 0.0095  max mem: 2905
Epoch: [100]  [400/625]  eta: 0:07:18  lr: 0.003236  min_lr: 0.003236  loss: 3.1646 (3.1680)  class_acc: 0.4922 (0.4901)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9940 (inf)  time: 1.7774  data: 0.0140  max mem: 2905
Epoch: [100]  [600/625]  eta: 0:00:48  lr: 0.003230  min_lr: 0.003230  loss: 3.1874 (3.1789)  class_acc: 0.4766 (0.4879)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.8117 (inf)  time: 1.8698  data: 0.0199  max mem: 2905
Epoch: [100]  [624/625]  eta: 0:00:01  lr: 0.003230  min_lr: 0.003230  loss: 3.1544 (3.1785)  class_acc: 0.4805 (0.4877)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9590 (inf)  time: 0.8299  data: 0.0016  max mem: 2905
Epoch: [100] Total time: 0:19:40 (1.8881 s / it)
Averaged stats: lr: 0.003230  min_lr: 0.003230  loss: 3.1544 (3.1749)  class_acc: 0.4805 (0.4877)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9590 (inf)
Test:  [ 0/50]  eta: 0:10:30  loss: 3.2758 (3.2758)  acc1: 32.0000 (32.0000)  acc5: 60.0000 (60.0000)  time: 12.6001  data: 12.5677  max mem: 2905
Test:  [10/50]  eta: 0:01:17  loss: 2.9321 (2.8179)  acc1: 41.6000 (41.5273)  acc5: 68.0000 (67.8545)  time: 1.9304  data: 1.9080  max mem: 2905
Test:  [20/50]  eta: 0:00:45  loss: 2.9604 (2.9723)  acc1: 37.6000 (38.7048)  acc5: 65.6000 (65.4476)  time: 0.9590  data: 0.9392  max mem: 2905
Test:  [30/50]  eta: 0:00:26  loss: 3.0349 (2.9716)  acc1: 37.6000 (39.6645)  acc5: 63.2000 (65.1871)  time: 0.9637  data: 0.9449  max mem: 2905
Test:  [40/50]  eta: 0:00:10  loss: 2.9051 (2.9577)  acc1: 40.8000 (39.9415)  acc5: 65.6000 (65.1317)  time: 0.5951  data: 0.5765  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 2.7267 (2.9271)  acc1: 40.8000 (39.8880)  acc5: 68.0000 (65.5840)  time: 0.5293  data: 0.5115  max mem: 2905
Test: Total time: 0:00:45 (0.9121 s / it)
* Acc@1 40.750 Acc@5 65.856 loss 2.890
Accuracy of the model on the 50000 test images: 40.8%
Max accuracy: 50.54%
Epoch: [101]  [  0/625]  eta: 3:33:06  lr: 0.003230  min_lr: 0.003230  loss: 3.1376 (3.1376)  class_acc: 0.5000 (0.5000)  weight_decay: 0.0500 (0.0500)  time: 20.4577  data: 20.0765  max mem: 2905
Epoch: [101]  [200/625]  eta: 0:14:02  lr: 0.003224  min_lr: 0.003224  loss: 3.1664 (3.1466)  class_acc: 0.4922 (0.4948)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7249 (2.4280)  time: 1.7851  data: 0.9487  max mem: 2905
Epoch: [101]  [400/625]  eta: 0:07:19  lr: 0.003218  min_lr: 0.003218  loss: 3.1468 (3.1597)  class_acc: 0.4961 (0.4907)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0668 (2.2974)  time: 1.9406  data: 0.0175  max mem: 2905
Epoch: [101]  [600/625]  eta: 0:00:49  lr: 0.003212  min_lr: 0.003212  loss: 3.1833 (3.1699)  class_acc: 0.4844 (0.4885)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4687 (2.4344)  time: 2.0241  data: 0.0007  max mem: 2905
Epoch: [101]  [624/625]  eta: 0:00:01  lr: 0.003212  min_lr: 0.003212  loss: 3.1917 (3.1714)  class_acc: 0.4688 (0.4881)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6580 (2.4374)  time: 0.7518  data: 0.0017  max mem: 2905
Epoch: [101] Total time: 0:19:59 (1.9193 s / it)
Averaged stats: lr: 0.003212  min_lr: 0.003212  loss: 3.1917 (3.1750)  class_acc: 0.4688 (0.4878)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6580 (2.4374)
Test:  [ 0/50]  eta: 0:10:25  loss: 2.9443 (2.9443)  acc1: 37.6000 (37.6000)  acc5: 65.6000 (65.6000)  time: 12.5021  data: 12.4728  max mem: 2905
Test:  [10/50]  eta: 0:01:27  loss: 2.9216 (2.9307)  acc1: 41.6000 (41.5273)  acc5: 64.8000 (64.5818)  time: 2.1930  data: 2.1736  max mem: 2905
Test:  [20/50]  eta: 0:00:53  loss: 3.0859 (3.1313)  acc1: 37.6000 (37.4857)  acc5: 63.2000 (61.6000)  time: 1.2344  data: 1.2157  max mem: 2905
Test:  [30/50]  eta: 0:00:31  loss: 3.1563 (3.1153)  acc1: 34.4000 (37.4968)  acc5: 60.8000 (61.8065)  time: 1.2435  data: 1.2252  max mem: 2905
Test:  [40/50]  eta: 0:00:13  loss: 3.1352 (3.1408)  acc1: 36.8000 (37.0341)  acc5: 60.8000 (61.7171)  time: 0.8088  data: 0.7905  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 3.2198 (3.1685)  acc1: 35.2000 (36.5760)  acc5: 60.0000 (61.2320)  time: 0.7152  data: 0.6960  max mem: 2905
Test: Total time: 0:00:54 (1.0895 s / it)
* Acc@1 37.208 Acc@5 62.040 loss 3.130
Accuracy of the model on the 50000 test images: 37.2%
Max accuracy: 50.54%
Epoch: [102]  [  0/625]  eta: 3:51:59  lr: 0.003212  min_lr: 0.003212  loss: 3.0261 (3.0261)  class_acc: 0.5273 (0.5273)  weight_decay: 0.0500 (0.0500)  time: 22.2708  data: 20.7913  max mem: 2905
Epoch: [102]  [200/625]  eta: 0:14:13  lr: 0.003206  min_lr: 0.003206  loss: 3.1967 (3.1444)  class_acc: 0.4883 (0.4942)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2358 (2.3092)  time: 1.8325  data: 1.3429  max mem: 2905
Epoch: [102]  [400/625]  eta: 0:07:29  lr: 0.003200  min_lr: 0.003200  loss: 3.2087 (3.1594)  class_acc: 0.4648 (0.4911)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3602 (2.2920)  time: 1.9981  data: 0.0009  max mem: 2905
Epoch: [102]  [600/625]  eta: 0:00:49  lr: 0.003195  min_lr: 0.003195  loss: 3.1908 (3.1655)  class_acc: 0.5000 (0.4896)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5207 (2.3051)  time: 2.0056  data: 0.0014  max mem: 2905
Epoch: [102]  [624/625]  eta: 0:00:01  lr: 0.003194  min_lr: 0.003194  loss: 3.2182 (3.1673)  class_acc: 0.4688 (0.4891)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5357 (2.3254)  time: 0.8929  data: 0.0032  max mem: 2905
Epoch: [102] Total time: 0:20:20 (1.9535 s / it)
Averaged stats: lr: 0.003194  min_lr: 0.003194  loss: 3.2182 (3.1719)  class_acc: 0.4688 (0.4884)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5357 (2.3254)
Test:  [ 0/50]  eta: 0:11:13  loss: 3.2653 (3.2653)  acc1: 35.2000 (35.2000)  acc5: 60.8000 (60.8000)  time: 13.4636  data: 13.4379  max mem: 2905
Test:  [10/50]  eta: 0:01:33  loss: 3.4063 (3.3121)  acc1: 32.8000 (34.1818)  acc5: 56.8000 (58.4727)  time: 2.3348  data: 2.3139  max mem: 2905
Test:  [20/50]  eta: 0:00:55  loss: 3.4072 (3.3944)  acc1: 32.8000 (33.0286)  acc5: 58.4000 (58.1714)  time: 1.2554  data: 1.2359  max mem: 2905
Test:  [30/50]  eta: 0:00:30  loss: 3.4000 (3.3726)  acc1: 34.4000 (34.2194)  acc5: 59.2000 (58.1677)  time: 1.0599  data: 1.0410  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 3.3375 (3.3585)  acc1: 35.2000 (34.3024)  acc5: 56.8000 (58.1659)  time: 0.6676  data: 0.6482  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 3.4321 (3.3680)  acc1: 33.6000 (34.0480)  acc5: 56.8000 (58.0480)  time: 0.5880  data: 0.5689  max mem: 2905
Test: Total time: 0:00:53 (1.0638 s / it)
* Acc@1 34.996 Acc@5 58.714 loss 3.336
Accuracy of the model on the 50000 test images: 35.0%
Max accuracy: 50.54%
Epoch: [103]  [  0/625]  eta: 3:43:39  lr: 0.003194  min_lr: 0.003194  loss: 3.0953 (3.0953)  class_acc: 0.5117 (0.5117)  weight_decay: 0.0500 (0.0500)  time: 21.4712  data: 19.3919  max mem: 2905
Epoch: [103]  [200/625]  eta: 0:14:27  lr: 0.003188  min_lr: 0.003188  loss: 3.2309 (3.1527)  class_acc: 0.4766 (0.4925)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.6373 (2.3989)  time: 1.8658  data: 0.0007  max mem: 2905
Epoch: [103]  [400/625]  eta: 0:07:25  lr: 0.003182  min_lr: 0.003182  loss: 3.1425 (3.1620)  class_acc: 0.4805 (0.4906)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7598 (2.3464)  time: 1.9866  data: 0.0007  max mem: 2905
Epoch: [103]  [600/625]  eta: 0:00:49  lr: 0.003176  min_lr: 0.003176  loss: 3.1077 (3.1660)  class_acc: 0.4883 (0.4895)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0958 (2.3843)  time: 1.9008  data: 0.0009  max mem: 2905
Epoch: [103]  [624/625]  eta: 0:00:01  lr: 0.003176  min_lr: 0.003176  loss: 3.1982 (3.1670)  class_acc: 0.4805 (0.4893)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6730 (2.3986)  time: 0.8285  data: 0.0019  max mem: 2905
Epoch: [103] Total time: 0:20:17 (1.9488 s / it)
Averaged stats: lr: 0.003176  min_lr: 0.003176  loss: 3.1982 (3.1690)  class_acc: 0.4805 (0.4892)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6730 (2.3986)
Test:  [ 0/50]  eta: 0:10:09  loss: 2.9171 (2.9171)  acc1: 38.4000 (38.4000)  acc5: 64.8000 (64.8000)  time: 12.1858  data: 12.1565  max mem: 2905
Test:  [10/50]  eta: 0:01:23  loss: 2.7624 (2.7191)  acc1: 45.6000 (45.0182)  acc5: 69.6000 (68.2182)  time: 2.0918  data: 2.0732  max mem: 2905
Test:  [20/50]  eta: 0:00:49  loss: 2.8051 (2.8198)  acc1: 40.8000 (42.3238)  acc5: 65.6000 (66.7810)  time: 1.1362  data: 1.1173  max mem: 2905
Test:  [30/50]  eta: 0:00:28  loss: 2.9191 (2.8257)  acc1: 39.2000 (42.0903)  acc5: 64.0000 (66.7355)  time: 1.0936  data: 1.0734  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 2.6934 (2.8243)  acc1: 40.8000 (42.3024)  acc5: 67.2000 (66.9854)  time: 0.7156  data: 0.6960  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 2.6455 (2.8337)  acc1: 41.6000 (42.0320)  acc5: 67.2000 (66.9280)  time: 0.7225  data: 0.7034  max mem: 2905
Test: Total time: 0:00:49 (0.9947 s / it)
* Acc@1 42.654 Acc@5 67.668 loss 2.807
Accuracy of the model on the 50000 test images: 42.7%
Max accuracy: 50.54%
Epoch: [104]  [  0/625]  eta: 3:38:52  lr: 0.003176  min_lr: 0.003176  loss: 3.3443 (3.3443)  class_acc: 0.4531 (0.4531)  weight_decay: 0.0500 (0.0500)  time: 21.0125  data: 20.8825  max mem: 2905
Epoch: [104]  [200/625]  eta: 0:14:28  lr: 0.003170  min_lr: 0.003170  loss: 3.1164 (3.1433)  class_acc: 0.5039 (0.4987)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5210 (2.5574)  time: 2.0468  data: 0.0043  max mem: 2905
Epoch: [104]  [400/625]  eta: 0:07:25  lr: 0.003164  min_lr: 0.003164  loss: 3.1662 (3.1526)  class_acc: 0.4805 (0.4945)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3693 (2.5750)  time: 2.0765  data: 0.0254  max mem: 2905
Epoch: [104]  [600/625]  eta: 0:00:49  lr: 0.003158  min_lr: 0.003158  loss: 3.2398 (3.1664)  class_acc: 0.4805 (0.4916)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1587 (2.4623)  time: 1.9450  data: 0.0044  max mem: 2905
Epoch: [104]  [624/625]  eta: 0:00:01  lr: 0.003158  min_lr: 0.003158  loss: 3.1528 (3.1661)  class_acc: 0.5000 (0.4917)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8106 (2.4427)  time: 0.8764  data: 0.0017  max mem: 2905
Epoch: [104] Total time: 0:20:02 (1.9237 s / it)
Averaged stats: lr: 0.003158  min_lr: 0.003158  loss: 3.1528 (3.1650)  class_acc: 0.5000 (0.4900)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8106 (2.4427)
Test:  [ 0/50]  eta: 0:09:13  loss: 2.9056 (2.9056)  acc1: 37.6000 (37.6000)  acc5: 70.4000 (70.4000)  time: 11.0689  data: 11.0371  max mem: 2905
Test:  [10/50]  eta: 0:01:16  loss: 2.7858 (2.7033)  acc1: 44.8000 (44.5091)  acc5: 70.4000 (68.2182)  time: 1.9054  data: 1.8855  max mem: 2905
Test:  [20/50]  eta: 0:00:45  loss: 2.8192 (2.8607)  acc1: 43.2000 (41.9048)  acc5: 64.8000 (66.3619)  time: 1.0375  data: 1.0188  max mem: 2905
Test:  [30/50]  eta: 0:00:27  loss: 2.9429 (2.8504)  acc1: 38.4000 (41.7548)  acc5: 64.8000 (66.6065)  time: 1.0493  data: 1.0301  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 2.9429 (2.8931)  acc1: 38.4000 (40.8000)  acc5: 65.6000 (66.1659)  time: 0.8328  data: 0.8139  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.9974 (2.8920)  acc1: 38.4000 (40.7520)  acc5: 64.0000 (66.2720)  time: 0.7642  data: 0.7457  max mem: 2905
Test: Total time: 0:00:51 (1.0282 s / it)
* Acc@1 41.038 Acc@5 66.228 loss 2.870
Accuracy of the model on the 50000 test images: 41.0%
Max accuracy: 50.54%
Epoch: [105]  [  0/625]  eta: 3:31:32  lr: 0.003158  min_lr: 0.003158  loss: 3.2014 (3.2014)  class_acc: 0.5000 (0.5000)  weight_decay: 0.0500 (0.0500)  time: 20.3073  data: 20.1807  max mem: 2905
Epoch: [105]  [200/625]  eta: 0:14:14  lr: 0.003152  min_lr: 0.003152  loss: 3.1316 (3.1406)  class_acc: 0.4961 (0.4958)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2005 (2.4949)  time: 1.9919  data: 1.1444  max mem: 2905
Epoch: [105]  [400/625]  eta: 0:07:23  lr: 0.003146  min_lr: 0.003146  loss: 3.1241 (3.1512)  class_acc: 0.4961 (0.4927)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2250 (2.3531)  time: 2.1765  data: 0.0260  max mem: 2905
Epoch: [105]  [600/625]  eta: 0:00:49  lr: 0.003140  min_lr: 0.003140  loss: 3.2256 (3.1601)  class_acc: 0.4727 (0.4906)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8167 (2.3134)  time: 2.0227  data: 0.0117  max mem: 2905
Epoch: [105]  [624/625]  eta: 0:00:01  lr: 0.003139  min_lr: 0.003139  loss: 3.1774 (3.1607)  class_acc: 0.4883 (0.4905)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0872 (2.3286)  time: 0.8843  data: 0.0110  max mem: 2905
Epoch: [105] Total time: 0:20:13 (1.9421 s / it)
Averaged stats: lr: 0.003139  min_lr: 0.003139  loss: 3.1774 (3.1608)  class_acc: 0.4883 (0.4910)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0872 (2.3286)
Test:  [ 0/50]  eta: 0:10:53  loss: 3.0985 (3.0985)  acc1: 38.4000 (38.4000)  acc5: 57.6000 (57.6000)  time: 13.0717  data: 13.0460  max mem: 2905
Test:  [10/50]  eta: 0:01:25  loss: 2.7367 (2.7793)  acc1: 42.4000 (42.3273)  acc5: 68.0000 (67.3455)  time: 2.1423  data: 2.1236  max mem: 2905
Test:  [20/50]  eta: 0:00:50  loss: 2.9104 (2.9556)  acc1: 39.2000 (39.1619)  acc5: 65.6000 (65.2191)  time: 1.1294  data: 1.1112  max mem: 2905
Test:  [30/50]  eta: 0:00:30  loss: 3.0784 (2.9826)  acc1: 36.8000 (38.8903)  acc5: 64.0000 (64.8774)  time: 1.1902  data: 1.1708  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 3.0541 (3.0222)  acc1: 36.8000 (38.3220)  acc5: 62.4000 (64.1366)  time: 0.8381  data: 0.8176  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 3.0541 (3.0259)  acc1: 37.6000 (38.4160)  acc5: 62.4000 (64.2240)  time: 0.8281  data: 0.8090  max mem: 2905
Test: Total time: 0:00:52 (1.0554 s / it)
* Acc@1 38.796 Acc@5 64.274 loss 2.991
Accuracy of the model on the 50000 test images: 38.8%
Max accuracy: 50.54%
Epoch: [106]  [  0/625]  eta: 3:26:28  lr: 0.003139  min_lr: 0.003139  loss: 3.0712 (3.0712)  class_acc: 0.5156 (0.5156)  weight_decay: 0.0500 (0.0500)  time: 19.8212  data: 16.6009  max mem: 2905
Epoch: [106]  [200/625]  eta: 0:14:34  lr: 0.003133  min_lr: 0.003133  loss: 3.0933 (3.1407)  class_acc: 0.5000 (0.4959)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9702 (2.2952)  time: 1.8725  data: 0.0009  max mem: 2905
Epoch: [106]  [400/625]  eta: 0:07:33  lr: 0.003127  min_lr: 0.003127  loss: 3.1157 (3.1477)  class_acc: 0.5039 (0.4948)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2686 (2.2832)  time: 1.9957  data: 0.0013  max mem: 2905
Epoch: [106]  [600/625]  eta: 0:00:50  lr: 0.003121  min_lr: 0.003121  loss: 3.1503 (3.1548)  class_acc: 0.4883 (0.4930)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1014 (inf)  time: 1.9333  data: 0.0008  max mem: 2905
Epoch: [106]  [624/625]  eta: 0:00:01  lr: 0.003121  min_lr: 0.003121  loss: 3.1527 (3.1560)  class_acc: 0.4805 (0.4928)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7173 (inf)  time: 0.7861  data: 0.0016  max mem: 2905
Epoch: [106] Total time: 0:20:23 (1.9584 s / it)
Averaged stats: lr: 0.003121  min_lr: 0.003121  loss: 3.1527 (3.1577)  class_acc: 0.4805 (0.4914)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7173 (inf)
Test:  [ 0/50]  eta: 0:10:23  loss: 2.6026 (2.6026)  acc1: 42.4000 (42.4000)  acc5: 73.6000 (73.6000)  time: 12.4737  data: 12.4500  max mem: 2905
Test:  [10/50]  eta: 0:01:22  loss: 2.6030 (2.6332)  acc1: 47.2000 (47.5636)  acc5: 72.0000 (70.1818)  time: 2.0535  data: 2.0331  max mem: 2905
Test:  [20/50]  eta: 0:00:48  loss: 2.6467 (2.7372)  acc1: 43.2000 (43.5810)  acc5: 68.8000 (68.4952)  time: 1.0799  data: 1.0594  max mem: 2905
Test:  [30/50]  eta: 0:00:28  loss: 2.8302 (2.7560)  acc1: 40.0000 (43.1226)  acc5: 67.2000 (67.9226)  time: 1.0699  data: 1.0501  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 2.8621 (2.7954)  acc1: 41.6000 (42.9268)  acc5: 67.2000 (67.3756)  time: 0.7944  data: 0.7759  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.8119 (2.8224)  acc1: 42.4000 (42.8960)  acc5: 66.4000 (66.9440)  time: 0.7183  data: 0.6997  max mem: 2905
Test: Total time: 0:00:51 (1.0240 s / it)
* Acc@1 43.326 Acc@5 68.176 loss 2.768
Accuracy of the model on the 50000 test images: 43.3%
Max accuracy: 50.54%
Epoch: [107]  [  0/625]  eta: 3:23:31  lr: 0.003121  min_lr: 0.003121  loss: 3.1516 (3.1516)  class_acc: 0.5430 (0.5430)  weight_decay: 0.0500 (0.0500)  time: 19.5378  data: 17.3284  max mem: 2905
Epoch: [107]  [200/625]  eta: 0:13:57  lr: 0.003115  min_lr: 0.003115  loss: 3.2130 (3.1431)  class_acc: 0.4766 (0.4946)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2384 (2.3360)  time: 1.6691  data: 0.8265  max mem: 2905
Epoch: [107]  [400/625]  eta: 0:07:18  lr: 0.003109  min_lr: 0.003109  loss: 3.1817 (3.1498)  class_acc: 0.4883 (0.4934)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0891 (2.3818)  time: 2.0108  data: 0.3347  max mem: 2905
Epoch: [107]  [600/625]  eta: 0:00:49  lr: 0.003103  min_lr: 0.003103  loss: 3.1289 (3.1537)  class_acc: 0.4922 (0.4930)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7743 (2.3908)  time: 2.0010  data: 0.0373  max mem: 2905
Epoch: [107]  [624/625]  eta: 0:00:01  lr: 0.003102  min_lr: 0.003102  loss: 3.1967 (3.1549)  class_acc: 0.4805 (0.4927)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.4697 (2.3685)  time: 0.8167  data: 0.0269  max mem: 2905
Epoch: [107] Total time: 0:20:02 (1.9234 s / it)
Averaged stats: lr: 0.003102  min_lr: 0.003102  loss: 3.1967 (3.1571)  class_acc: 0.4805 (0.4912)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.4697 (2.3685)
Test:  [ 0/50]  eta: 0:10:32  loss: 2.7467 (2.7467)  acc1: 37.6000 (37.6000)  acc5: 66.4000 (66.4000)  time: 12.6427  data: 12.5997  max mem: 2905
Test:  [10/50]  eta: 0:01:20  loss: 2.7949 (2.7865)  acc1: 44.8000 (43.9273)  acc5: 68.0000 (68.4364)  time: 2.0092  data: 1.9883  max mem: 2905
Test:  [20/50]  eta: 0:00:47  loss: 2.9228 (2.9221)  acc1: 39.2000 (39.9619)  acc5: 67.2000 (67.0476)  time: 1.0199  data: 1.0005  max mem: 2905
Test:  [30/50]  eta: 0:00:27  loss: 3.0798 (2.9494)  acc1: 36.8000 (40.2323)  acc5: 64.0000 (66.2710)  time: 1.0505  data: 1.0303  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 3.0080 (2.9402)  acc1: 40.8000 (40.4683)  acc5: 62.4000 (66.2634)  time: 0.8262  data: 0.8049  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.8797 (2.9281)  acc1: 40.8000 (40.6240)  acc5: 67.2000 (66.3680)  time: 0.7876  data: 0.7670  max mem: 2905
Test: Total time: 0:00:52 (1.0584 s / it)
* Acc@1 41.036 Acc@5 66.174 loss 2.896
Accuracy of the model on the 50000 test images: 41.0%
Max accuracy: 50.54%
Epoch: [108]  [  0/625]  eta: 3:38:43  lr: 0.003102  min_lr: 0.003102  loss: 3.1909 (3.1909)  class_acc: 0.4844 (0.4844)  weight_decay: 0.0500 (0.0500)  time: 20.9975  data: 16.6143  max mem: 2905
Epoch: [108]  [200/625]  eta: 0:14:29  lr: 0.003096  min_lr: 0.003096  loss: 3.1393 (3.1491)  class_acc: 0.4844 (0.4938)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8122 (2.2510)  time: 1.8571  data: 0.0008  max mem: 2905
Epoch: [108]  [400/625]  eta: 0:07:33  lr: 0.003090  min_lr: 0.003090  loss: 3.1372 (3.1548)  class_acc: 0.4805 (0.4926)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3752 (2.4471)  time: 1.9529  data: 0.0279  max mem: 2905
Epoch: [108]  [600/625]  eta: 0:00:50  lr: 0.003084  min_lr: 0.003084  loss: 3.1605 (3.1559)  class_acc: 0.4922 (0.4923)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.8046 (2.4023)  time: 1.7287  data: 0.0342  max mem: 2905
Epoch: [108]  [624/625]  eta: 0:00:01  lr: 0.003083  min_lr: 0.003083  loss: 3.1924 (3.1573)  class_acc: 0.4766 (0.4920)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0274 (2.3856)  time: 1.0585  data: 0.0261  max mem: 2905
Epoch: [108] Total time: 0:20:32 (1.9719 s / it)
Averaged stats: lr: 0.003083  min_lr: 0.003083  loss: 3.1924 (3.1517)  class_acc: 0.4766 (0.4928)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0274 (2.3856)
Test:  [ 0/50]  eta: 0:10:56  loss: 3.5392 (3.5392)  acc1: 26.4000 (26.4000)  acc5: 56.8000 (56.8000)  time: 13.1332  data: 13.1083  max mem: 2905
Test:  [10/50]  eta: 0:01:26  loss: 3.4241 (3.4111)  acc1: 33.6000 (33.1636)  acc5: 59.2000 (58.8364)  time: 2.1695  data: 2.1505  max mem: 2905
Test:  [20/50]  eta: 0:00:51  loss: 3.4225 (3.4542)  acc1: 32.8000 (32.4571)  acc5: 57.6000 (57.1810)  time: 1.1366  data: 1.1185  max mem: 2905
Test:  [30/50]  eta: 0:00:30  loss: 3.4147 (3.4114)  acc1: 33.6000 (33.8839)  acc5: 56.8000 (57.9097)  time: 1.1673  data: 1.1494  max mem: 2905
Test:  [40/50]  eta: 0:00:13  loss: 3.3768 (3.4268)  acc1: 36.0000 (33.6585)  acc5: 56.8000 (57.6390)  time: 0.9507  data: 0.9320  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 3.4960 (3.4551)  acc1: 32.8000 (33.2640)  acc5: 55.2000 (57.1680)  time: 0.8618  data: 0.8431  max mem: 2905
Test: Total time: 0:00:55 (1.1140 s / it)
* Acc@1 33.600 Acc@5 57.274 loss 3.427
Accuracy of the model on the 50000 test images: 33.6%
Max accuracy: 50.54%
Epoch: [109]  [  0/625]  eta: 4:20:55  lr: 0.003083  min_lr: 0.003083  loss: 3.1236 (3.1236)  class_acc: 0.5078 (0.5078)  weight_decay: 0.0500 (0.0500)  time: 25.0482  data: 20.8191  max mem: 2905
Epoch: [109]  [200/625]  eta: 0:15:00  lr: 0.003077  min_lr: 0.003077  loss: 3.1276 (3.1345)  class_acc: 0.4922 (0.4954)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5944 (2.2594)  time: 1.9012  data: 0.0019  max mem: 2905
Epoch: [109]  [400/625]  eta: 0:07:41  lr: 0.003071  min_lr: 0.003071  loss: 3.1508 (3.1457)  class_acc: 0.4922 (0.4936)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.6831 (2.2911)  time: 1.8350  data: 0.0012  max mem: 2905
Epoch: [109]  [600/625]  eta: 0:00:50  lr: 0.003065  min_lr: 0.003065  loss: 3.1741 (3.1498)  class_acc: 0.4766 (0.4923)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7819 (2.3950)  time: 2.0337  data: 0.0013  max mem: 2905
Epoch: [109]  [624/625]  eta: 0:00:01  lr: 0.003064  min_lr: 0.003064  loss: 3.1565 (3.1510)  class_acc: 0.4883 (0.4918)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8517 (2.3795)  time: 0.7991  data: 0.0015  max mem: 2905
Epoch: [109] Total time: 0:20:40 (1.9851 s / it)
Averaged stats: lr: 0.003064  min_lr: 0.003064  loss: 3.1565 (3.1535)  class_acc: 0.4883 (0.4926)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8517 (2.3795)
Test:  [ 0/50]  eta: 0:09:18  loss: 3.3506 (3.3506)  acc1: 26.4000 (26.4000)  acc5: 57.6000 (57.6000)  time: 11.1705  data: 11.1417  max mem: 2905
Test:  [10/50]  eta: 0:01:24  loss: 3.2450 (3.1655)  acc1: 38.4000 (37.2364)  acc5: 61.6000 (61.5273)  time: 2.1080  data: 2.0888  max mem: 2905
Test:  [20/50]  eta: 0:00:51  loss: 3.4150 (3.3660)  acc1: 35.2000 (33.5619)  acc5: 56.0000 (58.7429)  time: 1.2460  data: 1.2276  max mem: 2905
Test:  [30/50]  eta: 0:00:28  loss: 3.5671 (3.3309)  acc1: 32.0000 (34.0903)  acc5: 56.0000 (59.0710)  time: 1.0761  data: 1.0575  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 3.4460 (3.3640)  acc1: 34.4000 (34.1268)  acc5: 58.4000 (58.7707)  time: 0.6133  data: 0.5941  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 3.3549 (3.3470)  acc1: 34.4000 (34.4320)  acc5: 58.4000 (58.8640)  time: 0.4925  data: 0.4732  max mem: 2905
Test: Total time: 0:00:49 (0.9945 s / it)
* Acc@1 34.692 Acc@5 59.638 loss 3.313
Accuracy of the model on the 50000 test images: 34.7%
Max accuracy: 50.54%
Epoch: [110]  [  0/625]  eta: 3:27:59  lr: 0.003064  min_lr: 0.003064  loss: 3.2464 (3.2464)  class_acc: 0.4609 (0.4609)  weight_decay: 0.0500 (0.0500)  time: 19.9674  data: 18.5506  max mem: 2905
Epoch: [110]  [200/625]  eta: 0:14:01  lr: 0.003058  min_lr: 0.003058  loss: 3.1178 (3.1293)  class_acc: 0.5117 (0.4969)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5135 (2.3586)  time: 1.6940  data: 0.6570  max mem: 2905
Epoch: [110]  [400/625]  eta: 0:07:17  lr: 0.003052  min_lr: 0.003052  loss: 3.1297 (3.1314)  class_acc: 0.4883 (0.4965)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7539 (2.4547)  time: 1.9405  data: 0.0007  max mem: 2905
Epoch: [110]  [600/625]  eta: 0:00:48  lr: 0.003046  min_lr: 0.003046  loss: 3.1310 (3.1387)  class_acc: 0.4844 (0.4959)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0552 (2.4851)  time: 1.8424  data: 0.0008  max mem: 2905
Epoch: [110]  [624/625]  eta: 0:00:01  lr: 0.003045  min_lr: 0.003045  loss: 3.1845 (3.1403)  class_acc: 0.4844 (0.4954)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2768 (2.4909)  time: 0.6790  data: 0.0013  max mem: 2905
Epoch: [110] Total time: 0:20:02 (1.9243 s / it)
Averaged stats: lr: 0.003045  min_lr: 0.003045  loss: 3.1845 (3.1448)  class_acc: 0.4844 (0.4940)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2768 (2.4909)
Test:  [ 0/50]  eta: 0:10:39  loss: 3.3610 (3.3610)  acc1: 32.8000 (32.8000)  acc5: 56.0000 (56.0000)  time: 12.7838  data: 12.7582  max mem: 2905
Test:  [10/50]  eta: 0:01:23  loss: 3.3610 (3.4506)  acc1: 32.8000 (32.6545)  acc5: 56.0000 (57.1636)  time: 2.0904  data: 2.0713  max mem: 2905
Test:  [20/50]  eta: 0:00:48  loss: 3.5958 (3.6039)  acc1: 28.0000 (30.7048)  acc5: 52.8000 (54.5143)  time: 1.0588  data: 1.0394  max mem: 2905
Test:  [30/50]  eta: 0:00:27  loss: 3.5662 (3.5653)  acc1: 28.8000 (31.4065)  acc5: 53.6000 (55.0194)  time: 1.0114  data: 0.9916  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 3.4763 (3.5758)  acc1: 31.2000 (31.0634)  acc5: 54.4000 (54.8488)  time: 0.6688  data: 0.6495  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 3.4508 (3.5631)  acc1: 31.2000 (31.3280)  acc5: 57.6000 (55.2320)  time: 0.6077  data: 0.5879  max mem: 2905
Test: Total time: 0:00:48 (0.9734 s / it)
* Acc@1 32.310 Acc@5 56.064 loss 3.533
Accuracy of the model on the 50000 test images: 32.3%
Max accuracy: 50.54%
Epoch: [111]  [  0/625]  eta: 3:25:03  lr: 0.003045  min_lr: 0.003045  loss: 3.3081 (3.3081)  class_acc: 0.4609 (0.4609)  weight_decay: 0.0500 (0.0500)  time: 19.6849  data: 19.5456  max mem: 2905
Epoch: [111]  [200/625]  eta: 0:14:17  lr: 0.003039  min_lr: 0.003039  loss: 3.0880 (3.1228)  class_acc: 0.5117 (0.5016)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1752 (2.3099)  time: 1.7331  data: 0.0007  max mem: 2905
Epoch: [111]  [400/625]  eta: 0:07:20  lr: 0.003033  min_lr: 0.003033  loss: 3.1746 (3.1405)  class_acc: 0.4883 (0.4963)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4413 (2.3929)  time: 2.0589  data: 0.0402  max mem: 2905
Epoch: [111]  [600/625]  eta: 0:00:49  lr: 0.003027  min_lr: 0.003027  loss: 3.1327 (3.1485)  class_acc: 0.4922 (0.4945)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2550 (2.5366)  time: 2.0231  data: 0.0007  max mem: 2905
Epoch: [111]  [624/625]  eta: 0:00:01  lr: 0.003026  min_lr: 0.003026  loss: 3.1295 (3.1488)  class_acc: 0.4805 (0.4942)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1352 (2.5109)  time: 0.9173  data: 0.0012  max mem: 2905
Epoch: [111] Total time: 0:20:00 (1.9214 s / it)
Averaged stats: lr: 0.003026  min_lr: 0.003026  loss: 3.1295 (3.1461)  class_acc: 0.4805 (0.4943)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1352 (2.5109)
Test:  [ 0/50]  eta: 0:10:21  loss: 3.0910 (3.0910)  acc1: 38.4000 (38.4000)  acc5: 60.8000 (60.8000)  time: 12.4327  data: 12.4083  max mem: 2905
Test:  [10/50]  eta: 0:01:22  loss: 2.6630 (2.6809)  acc1: 44.0000 (44.2909)  acc5: 70.4000 (69.3091)  time: 2.0591  data: 2.0373  max mem: 2905
Test:  [20/50]  eta: 0:00:48  loss: 2.8214 (2.9004)  acc1: 41.6000 (40.9143)  acc5: 67.2000 (66.4381)  time: 1.0642  data: 1.0429  max mem: 2905
Test:  [30/50]  eta: 0:00:28  loss: 3.0150 (2.9705)  acc1: 36.0000 (40.0000)  acc5: 64.0000 (65.2129)  time: 1.0654  data: 1.0453  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 2.7833 (2.9176)  acc1: 41.6000 (40.8585)  acc5: 64.0000 (65.5220)  time: 0.8258  data: 0.8070  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.9047 (2.9343)  acc1: 36.8000 (40.2880)  acc5: 64.0000 (65.2000)  time: 0.7696  data: 0.7499  max mem: 2905
Test: Total time: 0:00:50 (1.0166 s / it)
* Acc@1 41.010 Acc@5 65.934 loss 2.893
Accuracy of the model on the 50000 test images: 41.0%
Max accuracy: 50.54%
Epoch: [112]  [  0/625]  eta: 3:35:00  lr: 0.003026  min_lr: 0.003026  loss: 2.9544 (2.9544)  class_acc: 0.5352 (0.5352)  weight_decay: 0.0500 (0.0500)  time: 20.6410  data: 18.8036  max mem: 2905
Epoch: [112]  [200/625]  eta: 0:14:23  lr: 0.003020  min_lr: 0.003020  loss: 3.1282 (3.1449)  class_acc: 0.4961 (0.4954)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8091 (2.3144)  time: 1.9008  data: 0.0387  max mem: 2905
Epoch: [112]  [400/625]  eta: 0:07:12  lr: 0.003014  min_lr: 0.003014  loss: 3.1076 (3.1472)  class_acc: 0.4844 (0.4954)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9001 (2.3687)  time: 1.6005  data: 0.0007  max mem: 2905
Epoch: [112]  [600/625]  eta: 0:00:47  lr: 0.003007  min_lr: 0.003007  loss: 3.1915 (3.1477)  class_acc: 0.4844 (0.4950)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2115 (2.3580)  time: 1.8564  data: 0.0006  max mem: 2905
Epoch: [112]  [624/625]  eta: 0:00:01  lr: 0.003007  min_lr: 0.003007  loss: 3.1449 (3.1481)  class_acc: 0.4961 (0.4949)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9117 (2.3456)  time: 0.7710  data: 0.0015  max mem: 2905
Epoch: [112] Total time: 0:19:40 (1.8893 s / it)
Averaged stats: lr: 0.003007  min_lr: 0.003007  loss: 3.1449 (3.1430)  class_acc: 0.4961 (0.4945)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9117 (2.3456)
Test:  [ 0/50]  eta: 0:09:59  loss: 3.3152 (3.3152)  acc1: 32.0000 (32.0000)  acc5: 60.0000 (60.0000)  time: 11.9977  data: 11.9675  max mem: 2905
Test:  [10/50]  eta: 0:01:20  loss: 3.0412 (3.0686)  acc1: 40.0000 (39.7818)  acc5: 62.4000 (63.5636)  time: 2.0132  data: 1.9925  max mem: 2905
Test:  [20/50]  eta: 0:00:47  loss: 3.1450 (3.2844)  acc1: 37.6000 (36.8381)  acc5: 60.0000 (60.8762)  time: 1.0595  data: 1.0393  max mem: 2905
Test:  [30/50]  eta: 0:00:26  loss: 3.2180 (3.2495)  acc1: 35.2000 (36.9806)  acc5: 58.4000 (60.8258)  time: 0.9231  data: 0.9036  max mem: 2905
Test:  [40/50]  eta: 0:00:10  loss: 3.2099 (3.2744)  acc1: 36.0000 (36.8000)  acc5: 57.6000 (60.5659)  time: 0.5613  data: 0.5431  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 3.3768 (3.2704)  acc1: 36.8000 (36.6400)  acc5: 57.6000 (60.7040)  time: 0.5332  data: 0.5143  max mem: 2905
Test: Total time: 0:00:46 (0.9208 s / it)
* Acc@1 36.796 Acc@5 61.040 loss 3.241
Accuracy of the model on the 50000 test images: 36.8%
Max accuracy: 50.54%
Epoch: [113]  [  0/625]  eta: 3:34:57  lr: 0.003007  min_lr: 0.003007  loss: 3.2775 (3.2775)  class_acc: 0.4727 (0.4727)  weight_decay: 0.0500 (0.0500)  time: 20.6364  data: 17.7362  max mem: 2905
Epoch: [113]  [200/625]  eta: 0:13:44  lr: 0.003000  min_lr: 0.003000  loss: 3.1433 (3.1208)  class_acc: 0.4961 (0.4994)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1387 (2.5876)  time: 1.8093  data: 0.0008  max mem: 2905
Epoch: [113]  [400/625]  eta: 0:07:12  lr: 0.002994  min_lr: 0.002994  loss: 3.1086 (3.1297)  class_acc: 0.4961 (0.4984)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1409 (inf)  time: 1.8178  data: 0.0008  max mem: 2905
Epoch: [113]  [600/625]  eta: 0:00:48  lr: 0.002988  min_lr: 0.002988  loss: 3.1627 (3.1366)  class_acc: 0.4922 (0.4969)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9516 (inf)  time: 2.0145  data: 0.0008  max mem: 2905
Epoch: [113]  [624/625]  eta: 0:00:01  lr: 0.002987  min_lr: 0.002987  loss: 3.1529 (3.1368)  class_acc: 0.4961 (0.4969)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2345 (inf)  time: 0.7779  data: 0.0015  max mem: 2905
Epoch: [113] Total time: 0:19:54 (1.9114 s / it)
Averaged stats: lr: 0.002987  min_lr: 0.002987  loss: 3.1529 (3.1396)  class_acc: 0.4961 (0.4951)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2345 (inf)
Test:  [ 0/50]  eta: 0:10:29  loss: 2.9229 (2.9229)  acc1: 37.6000 (37.6000)  acc5: 64.8000 (64.8000)  time: 12.5963  data: 12.5672  max mem: 2905
Test:  [10/50]  eta: 0:01:24  loss: 3.0741 (3.0143)  acc1: 39.2000 (39.0545)  acc5: 63.2000 (64.0727)  time: 2.1124  data: 2.0938  max mem: 2905
Test:  [20/50]  eta: 0:00:49  loss: 3.1591 (3.2334)  acc1: 35.2000 (35.8857)  acc5: 62.4000 (60.9905)  time: 1.1058  data: 1.0871  max mem: 2905
Test:  [30/50]  eta: 0:00:29  loss: 3.4525 (3.2071)  acc1: 33.6000 (36.6968)  acc5: 58.4000 (61.2645)  time: 1.1364  data: 1.1172  max mem: 2905
Test:  [40/50]  eta: 0:00:13  loss: 3.3710 (3.2426)  acc1: 36.8000 (36.3707)  acc5: 58.4000 (60.8390)  time: 1.0408  data: 1.0216  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 3.2432 (3.2648)  acc1: 36.8000 (35.9200)  acc5: 60.0000 (60.6400)  time: 0.9112  data: 0.8911  max mem: 2905
Test: Total time: 0:00:58 (1.1690 s / it)
* Acc@1 36.038 Acc@5 60.556 loss 3.247
Accuracy of the model on the 50000 test images: 36.0%
Max accuracy: 50.54%
Epoch: [114]  [  0/625]  eta: 3:33:10  lr: 0.002987  min_lr: 0.002987  loss: 3.3175 (3.3175)  class_acc: 0.4336 (0.4336)  weight_decay: 0.0500 (0.0500)  time: 20.4642  data: 16.8792  max mem: 2905
Epoch: [114]  [200/625]  eta: 0:14:40  lr: 0.002981  min_lr: 0.002981  loss: 3.1523 (3.1309)  class_acc: 0.5039 (0.4982)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1341 (2.2794)  time: 1.8696  data: 0.0009  max mem: 2905
Epoch: [114]  [400/625]  eta: 0:07:30  lr: 0.002975  min_lr: 0.002975  loss: 3.1316 (3.1342)  class_acc: 0.4961 (0.4968)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1540 (2.3170)  time: 1.9479  data: 0.0317  max mem: 2905
Epoch: [114]  [600/625]  eta: 0:00:49  lr: 0.002968  min_lr: 0.002968  loss: 3.1213 (3.1373)  class_acc: 0.5000 (0.4960)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1035 (2.4307)  time: 1.9822  data: 0.0335  max mem: 2905
Epoch: [114]  [624/625]  eta: 0:00:01  lr: 0.002968  min_lr: 0.002968  loss: 3.1687 (3.1388)  class_acc: 0.4766 (0.4956)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3021 (2.4420)  time: 0.8579  data: 0.0066  max mem: 2905
Epoch: [114] Total time: 0:20:08 (1.9328 s / it)
Averaged stats: lr: 0.002968  min_lr: 0.002968  loss: 3.1687 (3.1394)  class_acc: 0.4766 (0.4951)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3021 (2.4420)
Test:  [ 0/50]  eta: 0:09:14  loss: 3.5593 (3.5593)  acc1: 36.8000 (36.8000)  acc5: 54.4000 (54.4000)  time: 11.0843  data: 11.0517  max mem: 2905
Test:  [10/50]  eta: 0:01:22  loss: 3.1926 (3.2685)  acc1: 38.4000 (37.3818)  acc5: 60.0000 (58.9818)  time: 2.0692  data: 2.0476  max mem: 2905
Test:  [20/50]  eta: 0:00:49  loss: 3.3341 (3.4525)  acc1: 29.6000 (33.1810)  acc5: 57.6000 (57.3714)  time: 1.1825  data: 1.1630  max mem: 2905
Test:  [30/50]  eta: 0:00:28  loss: 3.6556 (3.4761)  acc1: 29.6000 (33.2129)  acc5: 56.8000 (57.1097)  time: 1.0764  data: 1.0585  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 3.5992 (3.5599)  acc1: 32.0000 (32.6049)  acc5: 56.8000 (56.0000)  time: 0.6618  data: 0.6445  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 3.7008 (3.5789)  acc1: 31.2000 (32.2400)  acc5: 52.8000 (55.8240)  time: 0.6254  data: 0.6076  max mem: 2905
Test: Total time: 0:00:48 (0.9657 s / it)
* Acc@1 32.702 Acc@5 56.232 loss 3.560
Accuracy of the model on the 50000 test images: 32.7%
Max accuracy: 50.54%
Epoch: [115]  [  0/625]  eta: 3:36:38  lr: 0.002968  min_lr: 0.002968  loss: 3.4248 (3.4248)  class_acc: 0.4531 (0.4531)  weight_decay: 0.0500 (0.0500)  time: 20.7979  data: 19.1501  max mem: 2905
Epoch: [115]  [200/625]  eta: 0:13:51  lr: 0.002961  min_lr: 0.002961  loss: 3.0897 (3.1254)  class_acc: 0.5000 (0.4982)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.6014 (2.2201)  time: 1.9206  data: 0.2626  max mem: 2905
Epoch: [115]  [400/625]  eta: 0:07:14  lr: 0.002955  min_lr: 0.002955  loss: 3.1286 (3.1279)  class_acc: 0.5000 (0.4982)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0474 (2.3980)  time: 1.9727  data: 0.0938  max mem: 2905
Epoch: [115]  [600/625]  eta: 0:00:48  lr: 0.002949  min_lr: 0.002949  loss: 3.1684 (3.1346)  class_acc: 0.4844 (0.4961)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0476 (2.5024)  time: 1.9562  data: 0.0008  max mem: 2905
Epoch: [115]  [624/625]  eta: 0:00:01  lr: 0.002948  min_lr: 0.002948  loss: 3.1492 (3.1352)  class_acc: 0.4961 (0.4961)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0476 (2.4942)  time: 0.7735  data: 0.0015  max mem: 2905
Epoch: [115] Total time: 0:19:50 (1.9051 s / it)
Averaged stats: lr: 0.002948  min_lr: 0.002948  loss: 3.1492 (3.1370)  class_acc: 0.4961 (0.4958)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0476 (2.4942)
Test:  [ 0/50]  eta: 0:09:49  loss: 3.2632 (3.2632)  acc1: 36.0000 (36.0000)  acc5: 62.4000 (62.4000)  time: 11.7911  data: 11.7616  max mem: 2905
Test:  [10/50]  eta: 0:01:19  loss: 2.8094 (2.8797)  acc1: 42.4000 (41.7455)  acc5: 64.8000 (65.8909)  time: 1.9933  data: 1.9718  max mem: 2905
Test:  [20/50]  eta: 0:00:46  loss: 2.8652 (2.9311)  acc1: 42.4000 (41.5619)  acc5: 64.8000 (65.2571)  time: 1.0551  data: 1.0359  max mem: 2905
Test:  [30/50]  eta: 0:00:28  loss: 2.8842 (2.9233)  acc1: 42.4000 (42.0129)  acc5: 66.4000 (65.1097)  time: 1.0874  data: 1.0695  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 2.8003 (2.9050)  acc1: 42.4000 (42.1073)  acc5: 67.2000 (65.7366)  time: 0.9582  data: 0.9403  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.8003 (2.9042)  acc1: 42.4000 (42.4640)  acc5: 67.2000 (65.7760)  time: 0.8729  data: 0.8522  max mem: 2905
Test: Total time: 0:00:53 (1.0693 s / it)
* Acc@1 42.494 Acc@5 66.976 loss 2.840
Accuracy of the model on the 50000 test images: 42.5%
Max accuracy: 50.54%
Epoch: [116]  [  0/625]  eta: 3:38:25  lr: 0.002948  min_lr: 0.002948  loss: 3.1878 (3.1878)  class_acc: 0.4844 (0.4844)  weight_decay: 0.0500 (0.0500)  time: 20.9682  data: 19.3178  max mem: 2905
Epoch: [116]  [200/625]  eta: 0:14:33  lr: 0.002942  min_lr: 0.002942  loss: 3.1119 (3.1229)  class_acc: 0.5000 (0.4999)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8911 (2.1056)  time: 1.9544  data: 0.0007  max mem: 2905
Epoch: [116]  [400/625]  eta: 0:07:24  lr: 0.002935  min_lr: 0.002935  loss: 3.0774 (3.1305)  class_acc: 0.4922 (0.4985)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8437 (2.2882)  time: 1.9125  data: 0.3040  max mem: 2905
Epoch: [116]  [600/625]  eta: 0:00:49  lr: 0.002929  min_lr: 0.002929  loss: 3.1168 (3.1329)  class_acc: 0.4961 (0.4968)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2304 (2.3359)  time: 1.8592  data: 0.0549  max mem: 2905
Epoch: [116]  [624/625]  eta: 0:00:01  lr: 0.002928  min_lr: 0.002928  loss: 3.1823 (3.1342)  class_acc: 0.4844 (0.4964)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3173 (2.3465)  time: 0.8612  data: 0.0239  max mem: 2905
Epoch: [116] Total time: 0:19:59 (1.9199 s / it)
Averaged stats: lr: 0.002928  min_lr: 0.002928  loss: 3.1823 (3.1324)  class_acc: 0.4844 (0.4966)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3173 (2.3465)
Test:  [ 0/50]  eta: 0:08:39  loss: 4.1406 (4.1406)  acc1: 28.8000 (28.8000)  acc5: 44.0000 (44.0000)  time: 10.3957  data: 10.3660  max mem: 2905
Test:  [10/50]  eta: 0:01:15  loss: 3.5837 (3.5091)  acc1: 34.4000 (33.0909)  acc5: 57.6000 (56.0000)  time: 1.8812  data: 1.8603  max mem: 2905
Test:  [20/50]  eta: 0:00:50  loss: 3.8511 (3.7375)  acc1: 28.0000 (29.4857)  acc5: 52.0000 (52.8762)  time: 1.2523  data: 1.2323  max mem: 2905
Test:  [30/50]  eta: 0:00:28  loss: 3.8628 (3.7050)  acc1: 28.0000 (29.5484)  acc5: 52.8000 (53.4194)  time: 1.1517  data: 1.1317  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 3.8783 (3.7928)  acc1: 27.2000 (28.7805)  acc5: 51.2000 (52.4878)  time: 0.6668  data: 0.6456  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 3.9732 (3.8051)  acc1: 25.6000 (28.6560)  acc5: 51.2000 (52.4640)  time: 0.4727  data: 0.4524  max mem: 2905
Test: Total time: 0:00:52 (1.0405 s / it)
* Acc@1 29.068 Acc@5 52.220 loss 3.771
Accuracy of the model on the 50000 test images: 29.1%
Max accuracy: 50.54%
Epoch: [117]  [  0/625]  eta: 3:28:43  lr: 0.002928  min_lr: 0.002928  loss: 3.0983 (3.0983)  class_acc: 0.5234 (0.5234)  weight_decay: 0.0500 (0.0500)  time: 20.0370  data: 19.3412  max mem: 2905
Epoch: [117]  [200/625]  eta: 0:14:07  lr: 0.002922  min_lr: 0.002922  loss: 3.1313 (3.1128)  class_acc: 0.5000 (0.5043)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9268 (2.4443)  time: 1.9684  data: 0.0007  max mem: 2905
Epoch: [117]  [400/625]  eta: 0:07:10  lr: 0.002915  min_lr: 0.002915  loss: 3.1125 (3.1223)  class_acc: 0.4922 (0.4998)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9851 (2.4194)  time: 1.9290  data: 0.0007  max mem: 2905
Epoch: [117]  [600/625]  eta: 0:00:47  lr: 0.002909  min_lr: 0.002909  loss: 3.2236 (3.1351)  class_acc: 0.4844 (0.4972)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7876 (2.4818)  time: 1.8074  data: 0.0006  max mem: 2905
Epoch: [117]  [624/625]  eta: 0:00:01  lr: 0.002908  min_lr: 0.002908  loss: 3.1674 (3.1353)  class_acc: 0.4922 (0.4971)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3374 (2.4869)  time: 0.9966  data: 0.0020  max mem: 2905
Epoch: [117] Total time: 0:19:42 (1.8925 s / it)
Averaged stats: lr: 0.002908  min_lr: 0.002908  loss: 3.1674 (3.1307)  class_acc: 0.4922 (0.4975)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3374 (2.4869)
Test:  [ 0/50]  eta: 0:11:42  loss: 3.9953 (3.9953)  acc1: 21.6000 (21.6000)  acc5: 48.0000 (48.0000)  time: 14.0416  data: 14.0124  max mem: 2905
Test:  [10/50]  eta: 0:01:39  loss: 4.0153 (4.0923)  acc1: 25.6000 (25.7455)  acc5: 47.2000 (47.2727)  time: 2.4778  data: 2.4549  max mem: 2905
Test:  [20/50]  eta: 0:00:57  loss: 4.1811 (4.1390)  acc1: 25.6000 (25.4857)  acc5: 47.2000 (46.8952)  time: 1.3048  data: 1.2837  max mem: 2905
Test:  [30/50]  eta: 0:00:32  loss: 4.1423 (4.0859)  acc1: 25.6000 (26.1677)  acc5: 48.0000 (48.2065)  time: 1.1654  data: 1.1449  max mem: 2905
Test:  [40/50]  eta: 0:00:13  loss: 4.1545 (4.1688)  acc1: 25.6000 (25.4634)  acc5: 44.8000 (47.2585)  time: 0.8025  data: 0.7828  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 4.1061 (4.1333)  acc1: 25.6000 (25.8240)  acc5: 46.4000 (47.6160)  time: 0.7169  data: 0.6953  max mem: 2905
Test: Total time: 0:00:56 (1.1385 s / it)
* Acc@1 26.426 Acc@5 47.990 loss 4.078
Accuracy of the model on the 50000 test images: 26.4%
Max accuracy: 50.54%
Epoch: [118]  [  0/625]  eta: 3:23:37  lr: 0.002908  min_lr: 0.002908  loss: 3.0896 (3.0896)  class_acc: 0.5039 (0.5039)  weight_decay: 0.0500 (0.0500)  time: 19.5480  data: 19.3674  max mem: 2905
Epoch: [118]  [200/625]  eta: 0:14:21  lr: 0.002902  min_lr: 0.002902  loss: 3.1041 (3.1083)  class_acc: 0.5117 (0.5010)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8109 (2.4007)  time: 1.9184  data: 0.0011  max mem: 2905
Epoch: [118]  [400/625]  eta: 0:07:22  lr: 0.002895  min_lr: 0.002895  loss: 3.0830 (3.1199)  class_acc: 0.5078 (0.4984)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9833 (2.3379)  time: 1.7695  data: 0.0008  max mem: 2905
Epoch: [118]  [600/625]  eta: 0:00:49  lr: 0.002889  min_lr: 0.002889  loss: 3.1148 (3.1254)  class_acc: 0.5039 (0.4970)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8102 (2.2762)  time: 1.8758  data: 0.0676  max mem: 2905
Epoch: [118]  [624/625]  eta: 0:00:01  lr: 0.002888  min_lr: 0.002888  loss: 3.0968 (3.1248)  class_acc: 0.4883 (0.4971)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8244 (2.3112)  time: 0.6965  data: 0.0173  max mem: 2905
Epoch: [118] Total time: 0:20:04 (1.9268 s / it)
Averaged stats: lr: 0.002888  min_lr: 0.002888  loss: 3.0968 (3.1292)  class_acc: 0.4883 (0.4974)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8244 (2.3112)
Test:  [ 0/50]  eta: 0:10:16  loss: 3.0749 (3.0749)  acc1: 38.4000 (38.4000)  acc5: 63.2000 (63.2000)  time: 12.3364  data: 12.3054  max mem: 2905
Test:  [10/50]  eta: 0:01:20  loss: 3.0426 (3.0186)  acc1: 38.4000 (39.1273)  acc5: 63.2000 (63.2727)  time: 2.0005  data: 1.9818  max mem: 2905
Test:  [20/50]  eta: 0:00:45  loss: 3.1473 (3.2063)  acc1: 35.2000 (35.8857)  acc5: 60.8000 (60.9143)  time: 0.9837  data: 0.9651  max mem: 2905
Test:  [30/50]  eta: 0:00:26  loss: 3.3094 (3.2278)  acc1: 33.6000 (35.7161)  acc5: 59.2000 (61.1613)  time: 0.9439  data: 0.9246  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 3.2905 (3.2561)  acc1: 33.6000 (35.5122)  acc5: 59.2000 (60.6244)  time: 0.7544  data: 0.7334  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 3.3631 (3.2862)  acc1: 33.6000 (35.0720)  acc5: 59.2000 (60.4960)  time: 0.7094  data: 0.6879  max mem: 2905
Test: Total time: 0:00:48 (0.9624 s / it)
* Acc@1 36.162 Acc@5 61.122 loss 3.242
Accuracy of the model on the 50000 test images: 36.2%
Max accuracy: 50.54%
Epoch: [119]  [  0/625]  eta: 3:32:10  lr: 0.002888  min_lr: 0.002888  loss: 3.1628 (3.1628)  class_acc: 0.4883 (0.4883)  weight_decay: 0.0500 (0.0500)  time: 20.3688  data: 18.4151  max mem: 2905
Epoch: [119]  [200/625]  eta: 0:14:32  lr: 0.002882  min_lr: 0.002882  loss: 3.1986 (3.1052)  class_acc: 0.4922 (0.5031)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9173 (2.2286)  time: 2.1279  data: 1.5632  max mem: 2905
Epoch: [119]  [400/625]  eta: 0:07:27  lr: 0.002875  min_lr: 0.002875  loss: 3.1299 (3.1126)  class_acc: 0.4883 (0.5006)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1469 (2.4205)  time: 1.8948  data: 1.7302  max mem: 2905
Epoch: [119]  [600/625]  eta: 0:00:49  lr: 0.002869  min_lr: 0.002869  loss: 3.1275 (3.1199)  class_acc: 0.4844 (0.4984)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5512 (inf)  time: 1.8978  data: 1.7007  max mem: 2905
Epoch: [119]  [624/625]  eta: 0:00:01  lr: 0.002868  min_lr: 0.002868  loss: 3.1683 (3.1216)  class_acc: 0.4922 (0.4982)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1216 (inf)  time: 0.7417  data: 0.5816  max mem: 2905
Epoch: [119] Total time: 0:20:02 (1.9246 s / it)
Averaged stats: lr: 0.002868  min_lr: 0.002868  loss: 3.1683 (3.1224)  class_acc: 0.4922 (0.4993)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1216 (inf)
Test:  [ 0/50]  eta: 0:09:59  loss: 4.2683 (4.2683)  acc1: 24.0000 (24.0000)  acc5: 44.0000 (44.0000)  time: 11.9993  data: 11.9667  max mem: 2905
Test:  [10/50]  eta: 0:01:26  loss: 3.5573 (3.6179)  acc1: 32.0000 (32.1455)  acc5: 55.2000 (54.9818)  time: 2.1551  data: 2.1356  max mem: 2905
Test:  [20/50]  eta: 0:00:50  loss: 3.8268 (3.7968)  acc1: 28.8000 (28.4952)  acc5: 52.8000 (52.8381)  time: 1.1803  data: 1.1608  max mem: 2905
Test:  [30/50]  eta: 0:00:27  loss: 3.8980 (3.8041)  acc1: 27.2000 (29.0839)  acc5: 52.0000 (52.9290)  time: 0.9089  data: 0.8884  max mem: 2905
Test:  [40/50]  eta: 0:00:10  loss: 3.7821 (3.7762)  acc1: 29.6000 (29.1512)  acc5: 52.0000 (53.2098)  time: 0.4686  data: 0.4494  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 3.6374 (3.7386)  acc1: 29.6000 (29.6000)  acc5: 54.4000 (53.6000)  time: 0.4130  data: 0.3949  max mem: 2905
Test: Total time: 0:00:46 (0.9211 s / it)
* Acc@1 30.378 Acc@5 54.038 loss 3.683
Accuracy of the model on the 50000 test images: 30.4%
Max accuracy: 50.54%
Epoch: [120]  [  0/625]  eta: 3:22:12  lr: 0.002868  min_lr: 0.002868  loss: 3.2221 (3.2221)  class_acc: 0.4883 (0.4883)  weight_decay: 0.0500 (0.0500)  time: 19.4124  data: 18.0886  max mem: 2905
Epoch: [120]  [200/625]  eta: 0:13:58  lr: 0.002862  min_lr: 0.002862  loss: 3.1484 (3.1049)  class_acc: 0.4961 (0.5015)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2431 (2.4987)  time: 1.8141  data: 0.0133  max mem: 2905
Epoch: [120]  [400/625]  eta: 0:07:12  lr: 0.002855  min_lr: 0.002855  loss: 3.1716 (3.1122)  class_acc: 0.4883 (0.5001)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.4651 (2.4838)  time: 1.9277  data: 0.0008  max mem: 2905
Epoch: [120]  [600/625]  eta: 0:00:47  lr: 0.002849  min_lr: 0.002849  loss: 3.1559 (3.1190)  class_acc: 0.4766 (0.4991)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9856 (2.4202)  time: 1.9562  data: 0.0239  max mem: 2905
Epoch: [120]  [624/625]  eta: 0:00:01  lr: 0.002848  min_lr: 0.002848  loss: 3.1083 (3.1197)  class_acc: 0.5039 (0.4989)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3195 (2.4117)  time: 0.7088  data: 0.0017  max mem: 2905
Epoch: [120] Total time: 0:19:28 (1.8694 s / it)
Averaged stats: lr: 0.002848  min_lr: 0.002848  loss: 3.1083 (3.1193)  class_acc: 0.5039 (0.4995)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3195 (2.4117)
Test:  [ 0/50]  eta: 0:10:14  loss: 3.2160 (3.2160)  acc1: 35.2000 (35.2000)  acc5: 58.4000 (58.4000)  time: 12.2955  data: 12.2578  max mem: 2905
Test:  [10/50]  eta: 0:01:16  loss: 3.2160 (3.2369)  acc1: 36.8000 (36.5091)  acc5: 59.2000 (59.9273)  time: 1.9078  data: 1.8871  max mem: 2905
Test:  [20/50]  eta: 0:00:42  loss: 3.4970 (3.4354)  acc1: 33.6000 (33.7143)  acc5: 57.6000 (57.3714)  time: 0.8834  data: 0.8641  max mem: 2905
Test:  [30/50]  eta: 0:00:26  loss: 3.5129 (3.4275)  acc1: 32.8000 (34.2710)  acc5: 52.8000 (57.6000)  time: 0.9770  data: 0.9563  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 3.4612 (3.4435)  acc1: 34.4000 (34.4390)  acc5: 58.4000 (57.8146)  time: 0.9414  data: 0.9208  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 3.4612 (3.4346)  acc1: 34.4000 (34.2400)  acc5: 58.4000 (58.0800)  time: 0.6356  data: 0.6153  max mem: 2905
Test: Total time: 0:00:53 (1.0685 s / it)
* Acc@1 35.024 Acc@5 58.718 loss 3.397
Accuracy of the model on the 50000 test images: 35.0%
Max accuracy: 50.54%
Epoch: [121]  [  0/625]  eta: 3:46:09  lr: 0.002848  min_lr: 0.002848  loss: 3.1343 (3.1343)  class_acc: 0.5273 (0.5273)  weight_decay: 0.0500 (0.0500)  time: 21.7105  data: 21.5833  max mem: 2905
Epoch: [121]  [200/625]  eta: 0:13:43  lr: 0.002841  min_lr: 0.002841  loss: 3.1020 (3.0979)  class_acc: 0.5000 (0.5024)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0525 (2.4709)  time: 1.8017  data: 0.0150  max mem: 2905
Epoch: [121]  [400/625]  eta: 0:07:16  lr: 0.002835  min_lr: 0.002835  loss: 3.0542 (3.1099)  class_acc: 0.5156 (0.5016)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0552 (2.5925)  time: 1.9216  data: 0.0010  max mem: 2905
Epoch: [121]  [600/625]  eta: 0:00:48  lr: 0.002828  min_lr: 0.002828  loss: 3.1089 (3.1144)  class_acc: 0.5000 (0.5011)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2338 (2.4597)  time: 2.0491  data: 0.0202  max mem: 2905
Epoch: [121]  [624/625]  eta: 0:00:01  lr: 0.002827  min_lr: 0.002827  loss: 3.0468 (3.1138)  class_acc: 0.5078 (0.5012)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9030 (2.4433)  time: 0.5510  data: 0.0064  max mem: 2905
Epoch: [121] Total time: 0:20:02 (1.9241 s / it)
Averaged stats: lr: 0.002827  min_lr: 0.002827  loss: 3.0468 (3.1194)  class_acc: 0.5078 (0.4998)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9030 (2.4433)
Test:  [ 0/50]  eta: 0:11:05  loss: 2.9186 (2.9186)  acc1: 40.0000 (40.0000)  acc5: 67.2000 (67.2000)  time: 13.3024  data: 13.2690  max mem: 2905
Test:  [10/50]  eta: 0:01:25  loss: 2.9789 (3.1045)  acc1: 40.0000 (38.9818)  acc5: 62.4000 (61.6727)  time: 2.1324  data: 2.1116  max mem: 2905
Test:  [20/50]  eta: 0:00:44  loss: 3.1822 (3.2055)  acc1: 36.8000 (36.8000)  acc5: 60.8000 (60.6476)  time: 0.8952  data: 0.8753  max mem: 2905
Test:  [30/50]  eta: 0:00:25  loss: 3.2810 (3.2318)  acc1: 36.0000 (37.0065)  acc5: 58.4000 (60.1032)  time: 0.8147  data: 0.7945  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 3.2810 (3.2511)  acc1: 33.6000 (36.4878)  acc5: 58.4000 (60.3707)  time: 0.9387  data: 0.9179  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 3.1007 (3.2328)  acc1: 38.4000 (36.5600)  acc5: 62.4000 (60.6240)  time: 0.7069  data: 0.6854  max mem: 2905
Test: Total time: 0:00:54 (1.0818 s / it)
* Acc@1 36.668 Acc@5 61.252 loss 3.199
Accuracy of the model on the 50000 test images: 36.7%
Max accuracy: 50.54%
Epoch: [122]  [  0/625]  eta: 3:49:48  lr: 0.002827  min_lr: 0.002827  loss: 3.1040 (3.1040)  class_acc: 0.5273 (0.5273)  weight_decay: 0.0500 (0.0500)  time: 22.0619  data: 19.6065  max mem: 2905
Epoch: [122]  [200/625]  eta: 0:14:54  lr: 0.002821  min_lr: 0.002821  loss: 3.1230 (3.1029)  class_acc: 0.5000 (0.5030)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.2911 (2.3558)  time: 2.0302  data: 0.0065  max mem: 2905
Epoch: [122]  [400/625]  eta: 0:07:46  lr: 0.002814  min_lr: 0.002814  loss: 3.1109 (3.1038)  class_acc: 0.4961 (0.5038)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9128 (2.4172)  time: 2.0333  data: 0.0168  max mem: 2905
Epoch: [122]  [600/625]  eta: 0:00:51  lr: 0.002808  min_lr: 0.002808  loss: 3.1263 (3.1124)  class_acc: 0.5039 (0.5025)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1427 (2.4045)  time: 2.0243  data: 0.0320  max mem: 2905
Epoch: [122]  [624/625]  eta: 0:00:02  lr: 0.002807  min_lr: 0.002807  loss: 3.1364 (3.1136)  class_acc: 0.4961 (0.5020)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8771 (2.4042)  time: 0.8710  data: 0.0129  max mem: 2905
Epoch: [122] Total time: 0:20:56 (2.0110 s / it)
Averaged stats: lr: 0.002807  min_lr: 0.002807  loss: 3.1364 (3.1195)  class_acc: 0.4961 (0.4998)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8771 (2.4042)
Test:  [ 0/50]  eta: 0:10:47  loss: 2.4031 (2.4031)  acc1: 55.2000 (55.2000)  acc5: 75.2000 (75.2000)  time: 12.9530  data: 12.9225  max mem: 2905
Test:  [10/50]  eta: 0:01:30  loss: 2.5733 (2.6236)  acc1: 45.6000 (45.3091)  acc5: 69.6000 (70.9818)  time: 2.2582  data: 2.2391  max mem: 2905
Test:  [20/50]  eta: 0:00:53  loss: 2.7623 (2.7106)  acc1: 42.4000 (43.6952)  acc5: 69.6000 (70.0190)  time: 1.2297  data: 1.2101  max mem: 2905
Test:  [30/50]  eta: 0:00:30  loss: 2.7293 (2.6982)  acc1: 42.4000 (44.0258)  acc5: 70.4000 (69.9097)  time: 1.1424  data: 1.1218  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 2.7293 (2.7471)  acc1: 43.2000 (43.6683)  acc5: 68.8000 (69.3659)  time: 0.6703  data: 0.6496  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.7873 (2.7637)  acc1: 43.2000 (43.5680)  acc5: 67.2000 (68.9920)  time: 0.6040  data: 0.5840  max mem: 2905
Test: Total time: 0:00:51 (1.0319 s / it)
* Acc@1 43.874 Acc@5 69.140 loss 2.704
Accuracy of the model on the 50000 test images: 43.9%
Max accuracy: 50.54%
Epoch: [123]  [  0/625]  eta: 3:45:48  lr: 0.002807  min_lr: 0.002807  loss: 3.0981 (3.0981)  class_acc: 0.5273 (0.5273)  weight_decay: 0.0500 (0.0500)  time: 21.6769  data: 19.5205  max mem: 2905
Epoch: [123]  [200/625]  eta: 0:13:57  lr: 0.002800  min_lr: 0.002800  loss: 3.1774 (3.0866)  class_acc: 0.4922 (0.5060)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2369 (2.5515)  time: 1.7764  data: 0.9033  max mem: 2905
Epoch: [123]  [400/625]  eta: 0:07:23  lr: 0.002794  min_lr: 0.002794  loss: 3.1246 (3.0986)  class_acc: 0.4883 (0.5041)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8960 (2.6429)  time: 1.9137  data: 0.0069  max mem: 2905
Epoch: [123]  [600/625]  eta: 0:00:49  lr: 0.002787  min_lr: 0.002787  loss: 3.0566 (3.1097)  class_acc: 0.5117 (0.5018)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4988 (2.6974)  time: 2.1439  data: 0.2413  max mem: 2905
Epoch: [123]  [624/625]  eta: 0:00:01  lr: 0.002786  min_lr: 0.002786  loss: 3.1921 (3.1112)  class_acc: 0.4922 (0.5012)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2753 (2.6823)  time: 0.3701  data: 0.0022  max mem: 2905
Epoch: [123] Total time: 0:20:23 (1.9578 s / it)
Averaged stats: lr: 0.002786  min_lr: 0.002786  loss: 3.1921 (3.1139)  class_acc: 0.4922 (0.5014)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2753 (2.6823)
Test:  [ 0/50]  eta: 0:11:26  loss: 3.3095 (3.3095)  acc1: 34.4000 (34.4000)  acc5: 60.8000 (60.8000)  time: 13.7278  data: 13.6841  max mem: 2905
Test:  [10/50]  eta: 0:01:28  loss: 3.2626 (3.3004)  acc1: 36.0000 (34.9091)  acc5: 57.6000 (59.4909)  time: 2.2241  data: 2.2037  max mem: 2905
Test:  [20/50]  eta: 0:00:52  loss: 3.2967 (3.3582)  acc1: 33.6000 (34.2095)  acc5: 56.8000 (58.7048)  time: 1.1608  data: 1.1428  max mem: 2905
Test:  [30/50]  eta: 0:00:31  loss: 3.3093 (3.3349)  acc1: 34.4000 (34.7355)  acc5: 56.8000 (59.2258)  time: 1.2176  data: 1.1994  max mem: 2905
Test:  [40/50]  eta: 0:00:13  loss: 3.2431 (3.3348)  acc1: 36.8000 (34.8683)  acc5: 58.4000 (59.1220)  time: 0.9768  data: 0.9574  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 3.2611 (3.3471)  acc1: 32.8000 (34.7840)  acc5: 59.2000 (59.1360)  time: 0.9445  data: 0.9251  max mem: 2905
Test: Total time: 0:00:56 (1.1373 s / it)
* Acc@1 34.816 Acc@5 59.766 loss 3.311
Accuracy of the model on the 50000 test images: 34.8%
Max accuracy: 50.54%
Epoch: [124]  [  0/625]  eta: 4:07:18  lr: 0.002786  min_lr: 0.002786  loss: 2.8637 (2.8637)  class_acc: 0.5469 (0.5469)  weight_decay: 0.0500 (0.0500)  time: 23.7417  data: 18.2486  max mem: 2905
Epoch: [124]  [200/625]  eta: 0:14:14  lr: 0.002780  min_lr: 0.002780  loss: 3.1413 (3.0922)  class_acc: 0.4883 (0.5041)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1518 (2.3834)  time: 1.8396  data: 0.0544  max mem: 2905
Epoch: [124]  [400/625]  eta: 0:07:29  lr: 0.002773  min_lr: 0.002773  loss: 3.1112 (3.1076)  class_acc: 0.5039 (0.5026)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4841 (2.4374)  time: 1.8704  data: 0.3943  max mem: 2905
Epoch: [124]  [600/625]  eta: 0:00:50  lr: 0.002766  min_lr: 0.002766  loss: 3.0802 (3.1127)  class_acc: 0.4961 (0.5011)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0208 (2.3702)  time: 2.0075  data: 0.0648  max mem: 2905
Epoch: [124]  [624/625]  eta: 0:00:01  lr: 0.002766  min_lr: 0.002766  loss: 3.1260 (3.1140)  class_acc: 0.4961 (0.5007)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0441 (2.3610)  time: 0.7640  data: 0.0173  max mem: 2905
Epoch: [124] Total time: 0:20:29 (1.9671 s / it)
Averaged stats: lr: 0.002766  min_lr: 0.002766  loss: 3.1260 (3.1093)  class_acc: 0.4961 (0.5019)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0441 (2.3610)
Test:  [ 0/50]  eta: 0:10:06  loss: 3.5939 (3.5939)  acc1: 32.8000 (32.8000)  acc5: 52.8000 (52.8000)  time: 12.1348  data: 12.1105  max mem: 2905
Test:  [10/50]  eta: 0:01:18  loss: 3.0858 (3.1422)  acc1: 37.6000 (38.1818)  acc5: 64.0000 (62.8364)  time: 1.9732  data: 1.9545  max mem: 2905
Test:  [20/50]  eta: 0:00:46  loss: 3.2235 (3.3061)  acc1: 34.4000 (35.1619)  acc5: 57.6000 (59.4286)  time: 1.0053  data: 0.9864  max mem: 2905
Test:  [30/50]  eta: 0:00:26  loss: 3.4119 (3.3379)  acc1: 30.4000 (34.7355)  acc5: 56.8000 (59.2516)  time: 0.9775  data: 0.9579  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 3.3754 (3.3646)  acc1: 32.8000 (34.1463)  acc5: 58.4000 (58.9854)  time: 0.6632  data: 0.6438  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 3.3754 (3.3862)  acc1: 32.8000 (34.0640)  acc5: 57.6000 (58.6080)  time: 0.5680  data: 0.5491  max mem: 2905
Test: Total time: 0:00:47 (0.9570 s / it)
* Acc@1 34.428 Acc@5 58.544 loss 3.367
Accuracy of the model on the 50000 test images: 34.4%
Max accuracy: 50.54%
Epoch: [125]  [  0/625]  eta: 3:23:15  lr: 0.002766  min_lr: 0.002766  loss: 3.1610 (3.1610)  class_acc: 0.4844 (0.4844)  weight_decay: 0.0500 (0.0500)  time: 19.5122  data: 16.7370  max mem: 2905
Epoch: [125]  [200/625]  eta: 0:13:57  lr: 0.002759  min_lr: 0.002759  loss: 3.0773 (3.0800)  class_acc: 0.5000 (0.5088)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.5427 (2.4667)  time: 1.7531  data: 0.0006  max mem: 2905
Epoch: [125]  [400/625]  eta: 0:07:20  lr: 0.002752  min_lr: 0.002752  loss: 3.1404 (3.0988)  class_acc: 0.4922 (0.5046)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1039 (2.5865)  time: 2.0188  data: 0.0008  max mem: 2905
Epoch: [125]  [600/625]  eta: 0:00:49  lr: 0.002746  min_lr: 0.002746  loss: 3.1789 (3.1080)  class_acc: 0.4922 (0.5020)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3566 (2.5312)  time: 2.1738  data: 0.0006  max mem: 2905
Epoch: [125]  [624/625]  eta: 0:00:01  lr: 0.002745  min_lr: 0.002745  loss: 3.0558 (3.1074)  class_acc: 0.5039 (0.5021)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6566 (2.5402)  time: 0.6747  data: 0.0013  max mem: 2905
Epoch: [125] Total time: 0:20:05 (1.9282 s / it)
Averaged stats: lr: 0.002745  min_lr: 0.002745  loss: 3.0558 (3.1072)  class_acc: 0.5039 (0.5025)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6566 (2.5402)
Test:  [ 0/50]  eta: 0:09:55  loss: 2.9795 (2.9795)  acc1: 38.4000 (38.4000)  acc5: 64.0000 (64.0000)  time: 11.9024  data: 11.8643  max mem: 2905
Test:  [10/50]  eta: 0:01:25  loss: 2.8623 (2.8837)  acc1: 40.0000 (41.0182)  acc5: 64.8000 (65.1636)  time: 2.1405  data: 2.1095  max mem: 2905
Test:  [20/50]  eta: 0:00:52  loss: 2.9883 (3.1113)  acc1: 39.2000 (37.8286)  acc5: 62.4000 (62.2476)  time: 1.2308  data: 1.2063  max mem: 2905
Test:  [30/50]  eta: 0:00:31  loss: 3.2644 (3.1340)  acc1: 36.8000 (37.4968)  acc5: 60.8000 (62.0645)  time: 1.2606  data: 1.2421  max mem: 2905
Test:  [40/50]  eta: 0:00:13  loss: 3.2380 (3.1791)  acc1: 36.0000 (37.2293)  acc5: 60.0000 (61.6390)  time: 0.8953  data: 0.8763  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 3.2380 (3.1937)  acc1: 36.0000 (37.1840)  acc5: 60.0000 (61.4400)  time: 0.7498  data: 0.7303  max mem: 2905
Test: Total time: 0:00:56 (1.1222 s / it)
* Acc@1 37.044 Acc@5 62.218 loss 3.153
Accuracy of the model on the 50000 test images: 37.0%
Max accuracy: 50.54%
Epoch: [126]  [  0/625]  eta: 3:38:53  lr: 0.002745  min_lr: 0.002745  loss: 3.1257 (3.1257)  class_acc: 0.5039 (0.5039)  weight_decay: 0.0500 (0.0500)  time: 21.0129  data: 17.3480  max mem: 2905
Epoch: [126]  [200/625]  eta: 0:14:48  lr: 0.002738  min_lr: 0.002738  loss: 3.0536 (3.0802)  class_acc: 0.5039 (0.5091)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.0856 (2.4246)  time: 1.9752  data: 0.0040  max mem: 2905
Epoch: [126]  [400/625]  eta: 0:07:39  lr: 0.002732  min_lr: 0.002732  loss: 3.0944 (3.0932)  class_acc: 0.5078 (0.5060)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.6031 (inf)  time: 1.9729  data: 0.0014  max mem: 2905
Epoch: [126]  [600/625]  eta: 0:00:50  lr: 0.002725  min_lr: 0.002725  loss: 3.1125 (3.1016)  class_acc: 0.4961 (0.5040)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7345 (inf)  time: 1.9805  data: 0.0009  max mem: 2905
Epoch: [126]  [624/625]  eta: 0:00:01  lr: 0.002724  min_lr: 0.002724  loss: 3.1249 (3.1032)  class_acc: 0.4961 (0.5036)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0445 (inf)  time: 0.8589  data: 0.0016  max mem: 2905
Epoch: [126] Total time: 0:20:40 (1.9842 s / it)
Averaged stats: lr: 0.002724  min_lr: 0.002724  loss: 3.1249 (3.1052)  class_acc: 0.4961 (0.5032)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0445 (inf)
Test:  [ 0/50]  eta: 0:10:02  loss: 3.7799 (3.7799)  acc1: 27.2000 (27.2000)  acc5: 55.2000 (55.2000)  time: 12.0437  data: 12.0108  max mem: 2905
Test:  [10/50]  eta: 0:01:18  loss: 3.3612 (3.2897)  acc1: 37.6000 (37.3091)  acc5: 58.4000 (59.5636)  time: 1.9675  data: 1.9478  max mem: 2905
Test:  [20/50]  eta: 0:00:45  loss: 3.3612 (3.4181)  acc1: 35.2000 (35.0476)  acc5: 57.6000 (57.5238)  time: 1.0001  data: 0.9815  max mem: 2905
Test:  [30/50]  eta: 0:00:26  loss: 3.4702 (3.4304)  acc1: 33.6000 (34.7355)  acc5: 56.0000 (57.4710)  time: 0.9542  data: 0.9352  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 3.4367 (3.4353)  acc1: 32.8000 (34.4976)  acc5: 56.0000 (57.4244)  time: 0.6697  data: 0.6514  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 3.4229 (3.4391)  acc1: 32.0000 (34.5120)  acc5: 56.8000 (57.4080)  time: 0.6718  data: 0.6541  max mem: 2905
Test: Total time: 0:00:47 (0.9592 s / it)
* Acc@1 35.066 Acc@5 58.574 loss 3.400
Accuracy of the model on the 50000 test images: 35.1%
Max accuracy: 50.54%
Epoch: [127]  [  0/625]  eta: 3:23:21  lr: 0.002724  min_lr: 0.002724  loss: 3.0901 (3.0901)  class_acc: 0.5039 (0.5039)  weight_decay: 0.0500 (0.0500)  time: 19.5219  data: 19.4014  max mem: 2905
Epoch: [127]  [200/625]  eta: 0:14:34  lr: 0.002717  min_lr: 0.002717  loss: 3.0676 (3.0913)  class_acc: 0.5039 (0.5047)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2467 (2.4505)  time: 2.0436  data: 0.0007  max mem: 2905
Epoch: [127]  [400/625]  eta: 0:07:30  lr: 0.002711  min_lr: 0.002711  loss: 3.0939 (3.1025)  class_acc: 0.5156 (0.5019)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7163 (2.3695)  time: 1.8498  data: 0.0006  max mem: 2905
Epoch: [127]  [600/625]  eta: 0:00:49  lr: 0.002704  min_lr: 0.002704  loss: 3.1401 (3.1065)  class_acc: 0.4883 (0.5011)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8148 (2.4658)  time: 2.1132  data: 0.0006  max mem: 2905
Epoch: [127]  [624/625]  eta: 0:00:01  lr: 0.002703  min_lr: 0.002703  loss: 3.0607 (3.1061)  class_acc: 0.5117 (0.5012)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.6179 (2.4430)  time: 0.7911  data: 0.0019  max mem: 2905
Epoch: [127] Total time: 0:20:12 (1.9402 s / it)
Averaged stats: lr: 0.002703  min_lr: 0.002703  loss: 3.0607 (3.1065)  class_acc: 0.5117 (0.5024)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.6179 (2.4430)
Test:  [ 0/50]  eta: 0:10:37  loss: 3.6693 (3.6693)  acc1: 27.2000 (27.2000)  acc5: 52.8000 (52.8000)  time: 12.7422  data: 12.7054  max mem: 2905
Test:  [10/50]  eta: 0:01:23  loss: 3.2945 (3.1750)  acc1: 36.0000 (37.0182)  acc5: 60.0000 (60.5091)  time: 2.0804  data: 2.0589  max mem: 2905
Test:  [20/50]  eta: 0:00:48  loss: 3.3121 (3.4147)  acc1: 35.2000 (33.3714)  acc5: 58.4000 (57.1810)  time: 1.0583  data: 1.0391  max mem: 2905
Test:  [30/50]  eta: 0:00:29  loss: 3.5898 (3.4077)  acc1: 30.4000 (33.5742)  acc5: 53.6000 (57.1097)  time: 1.1034  data: 1.0851  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 3.3769 (3.4460)  acc1: 32.8000 (33.1317)  acc5: 57.6000 (56.9561)  time: 0.9683  data: 0.9500  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 3.4949 (3.4829)  acc1: 30.4000 (32.6400)  acc5: 55.2000 (56.3680)  time: 0.8643  data: 0.8457  max mem: 2905
Test: Total time: 0:00:56 (1.1320 s / it)
* Acc@1 33.440 Acc@5 57.588 loss 3.434
Accuracy of the model on the 50000 test images: 33.4%
Max accuracy: 50.54%
Epoch: [128]  [  0/625]  eta: 3:44:39  lr: 0.002703  min_lr: 0.002703  loss: 3.1353 (3.1353)  class_acc: 0.4531 (0.4531)  weight_decay: 0.0500 (0.0500)  time: 21.5670  data: 19.3903  max mem: 2905
Epoch: [128]  [200/625]  eta: 0:14:31  lr: 0.002696  min_lr: 0.002696  loss: 3.1144 (3.0818)  class_acc: 0.4922 (0.5080)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7458 (2.5138)  time: 2.1050  data: 0.8604  max mem: 2905
Epoch: [128]  [400/625]  eta: 0:07:29  lr: 0.002690  min_lr: 0.002690  loss: 3.0982 (3.0897)  class_acc: 0.5000 (0.5065)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3505 (2.4784)  time: 1.9242  data: 0.1620  max mem: 2905
Epoch: [128]  [600/625]  eta: 0:00:49  lr: 0.002683  min_lr: 0.002683  loss: 3.1106 (3.0972)  class_acc: 0.5039 (0.5052)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0343 (2.5318)  time: 1.9149  data: 0.0044  max mem: 2905
Epoch: [128]  [624/625]  eta: 0:00:01  lr: 0.002682  min_lr: 0.002682  loss: 3.1843 (3.0991)  class_acc: 0.4883 (0.5047)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9398 (2.5042)  time: 0.7570  data: 0.0106  max mem: 2905
Epoch: [128] Total time: 0:20:14 (1.9432 s / it)
Averaged stats: lr: 0.002682  min_lr: 0.002682  loss: 3.1843 (3.0996)  class_acc: 0.4883 (0.5041)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9398 (2.5042)
Test:  [ 0/50]  eta: 0:10:17  loss: 3.1274 (3.1274)  acc1: 38.4000 (38.4000)  acc5: 60.0000 (60.0000)  time: 12.3595  data: 12.3347  max mem: 2905
Test:  [10/50]  eta: 0:01:26  loss: 2.6262 (2.6298)  acc1: 46.4000 (46.0364)  acc5: 71.2000 (70.1818)  time: 2.1689  data: 2.1489  max mem: 2905
Test:  [20/50]  eta: 0:00:52  loss: 2.7527 (2.7734)  acc1: 41.6000 (42.3238)  acc5: 67.2000 (68.2286)  time: 1.2200  data: 1.2005  max mem: 2905
Test:  [30/50]  eta: 0:00:29  loss: 2.9066 (2.8009)  acc1: 41.6000 (42.6839)  acc5: 64.8000 (67.7936)  time: 1.0967  data: 1.0756  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 2.9212 (2.8277)  acc1: 41.6000 (42.1073)  acc5: 64.8000 (67.3756)  time: 0.6433  data: 0.6215  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.8017 (2.8137)  acc1: 41.6000 (42.2880)  acc5: 65.6000 (67.5360)  time: 0.5492  data: 0.5301  max mem: 2905
Test: Total time: 0:00:50 (1.0037 s / it)
* Acc@1 42.296 Acc@5 67.754 loss 2.787
Accuracy of the model on the 50000 test images: 42.3%
Max accuracy: 50.54%
Epoch: [129]  [  0/625]  eta: 3:45:57  lr: 0.002682  min_lr: 0.002682  loss: 3.0994 (3.0994)  class_acc: 0.4922 (0.4922)  weight_decay: 0.0500 (0.0500)  time: 21.6926  data: 15.0788  max mem: 2905
Epoch: [129]  [200/625]  eta: 0:13:54  lr: 0.002675  min_lr: 0.002675  loss: 3.1086 (3.0957)  class_acc: 0.5000 (0.5059)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1434 (2.2222)  time: 1.7782  data: 0.0087  max mem: 2905
Epoch: [129]  [400/625]  eta: 0:07:12  lr: 0.002668  min_lr: 0.002668  loss: 3.1449 (3.1054)  class_acc: 0.5000 (0.5021)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9761 (2.3517)  time: 1.9512  data: 0.0006  max mem: 2905
Epoch: [129]  [600/625]  eta: 0:00:48  lr: 0.002662  min_lr: 0.002662  loss: 3.1047 (3.1075)  class_acc: 0.4961 (0.5017)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2787 (2.3888)  time: 2.1130  data: 0.0007  max mem: 2905
Epoch: [129]  [624/625]  eta: 0:00:01  lr: 0.002661  min_lr: 0.002661  loss: 3.1605 (3.1092)  class_acc: 0.5000 (0.5015)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8851 (2.4000)  time: 0.7090  data: 0.0018  max mem: 2905
Epoch: [129] Total time: 0:20:05 (1.9283 s / it)
Averaged stats: lr: 0.002661  min_lr: 0.002661  loss: 3.1605 (3.0982)  class_acc: 0.5000 (0.5045)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8851 (2.4000)
Test:  [ 0/50]  eta: 0:10:03  loss: 2.2711 (2.2711)  acc1: 52.0000 (52.0000)  acc5: 76.0000 (76.0000)  time: 12.0701  data: 12.0443  max mem: 2905
Test:  [10/50]  eta: 0:01:16  loss: 2.5456 (2.5777)  acc1: 45.6000 (46.6909)  acc5: 73.6000 (70.1818)  time: 1.9042  data: 1.8842  max mem: 2905
Test:  [20/50]  eta: 0:00:44  loss: 2.7482 (2.6799)  acc1: 43.2000 (44.1524)  acc5: 66.4000 (68.9143)  time: 0.9414  data: 0.9204  max mem: 2905
Test:  [30/50]  eta: 0:00:25  loss: 2.8096 (2.7056)  acc1: 42.4000 (44.0258)  acc5: 65.6000 (68.6710)  time: 0.9571  data: 0.9368  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 2.8800 (2.7397)  acc1: 42.4000 (43.6488)  acc5: 67.2000 (68.4488)  time: 0.9694  data: 0.9504  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.8305 (2.7510)  acc1: 42.4000 (43.4720)  acc5: 67.2000 (68.3520)  time: 0.9179  data: 0.8981  max mem: 2905
Test: Total time: 0:00:55 (1.1118 s / it)
* Acc@1 44.050 Acc@5 69.110 loss 2.714
Accuracy of the model on the 50000 test images: 44.1%
Max accuracy: 50.54%
Epoch: [130]  [  0/625]  eta: 3:42:07  lr: 0.002661  min_lr: 0.002661  loss: 2.9074 (2.9074)  class_acc: 0.5508 (0.5508)  weight_decay: 0.0500 (0.0500)  time: 21.3248  data: 20.4725  max mem: 2905
Epoch: [130]  [200/625]  eta: 0:14:48  lr: 0.002654  min_lr: 0.002654  loss: 3.0971 (3.0695)  class_acc: 0.5000 (0.5100)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.7519 (2.7999)  time: 2.0013  data: 0.0042  max mem: 2905
Epoch: [130]  [400/625]  eta: 0:07:36  lr: 0.002647  min_lr: 0.002647  loss: 3.0981 (3.0921)  class_acc: 0.5000 (0.5066)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8191 (2.6758)  time: 2.0823  data: 0.0007  max mem: 2905
Epoch: [130]  [600/625]  eta: 0:00:50  lr: 0.002640  min_lr: 0.002640  loss: 3.0461 (3.0995)  class_acc: 0.5039 (0.5051)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4477 (2.5700)  time: 2.0177  data: 0.0009  max mem: 2905
Epoch: [130]  [624/625]  eta: 0:00:01  lr: 0.002640  min_lr: 0.002640  loss: 3.0996 (3.1000)  class_acc: 0.4922 (0.5051)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4012 (2.5652)  time: 0.7370  data: 0.0013  max mem: 2905
Epoch: [130] Total time: 0:20:27 (1.9640 s / it)
Averaged stats: lr: 0.002640  min_lr: 0.002640  loss: 3.0996 (3.0959)  class_acc: 0.4922 (0.5052)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4012 (2.5652)
Test:  [ 0/50]  eta: 0:09:59  loss: 2.6034 (2.6034)  acc1: 43.2000 (43.2000)  acc5: 75.2000 (75.2000)  time: 11.9956  data: 11.9550  max mem: 2905
Test:  [10/50]  eta: 0:01:24  loss: 2.6750 (2.7373)  acc1: 44.0000 (42.9818)  acc5: 66.4000 (67.2000)  time: 2.1179  data: 2.0949  max mem: 2905
Test:  [20/50]  eta: 0:00:50  loss: 2.8872 (2.8578)  acc1: 40.0000 (40.4952)  acc5: 64.8000 (65.6000)  time: 1.1698  data: 1.1492  max mem: 2905
Test:  [30/50]  eta: 0:00:27  loss: 2.9208 (2.8683)  acc1: 38.4000 (40.8774)  acc5: 64.0000 (65.4968)  time: 0.9990  data: 0.9794  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 2.9052 (2.8803)  acc1: 40.8000 (40.7805)  acc5: 64.8000 (65.5220)  time: 0.5554  data: 0.5356  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 2.8313 (2.8828)  acc1: 40.8000 (40.8160)  acc5: 64.8000 (65.8720)  time: 0.4990  data: 0.4795  max mem: 2905
Test: Total time: 0:00:46 (0.9368 s / it)
* Acc@1 41.654 Acc@5 66.666 loss 2.834
Accuracy of the model on the 50000 test images: 41.7%
Max accuracy: 50.54%
Epoch: [131]  [  0/625]  eta: 3:48:14  lr: 0.002640  min_lr: 0.002640  loss: 3.0310 (3.0310)  class_acc: 0.5156 (0.5156)  weight_decay: 0.0500 (0.0500)  time: 21.9107  data: 16.7253  max mem: 2905
Epoch: [131]  [200/625]  eta: 0:14:34  lr: 0.002633  min_lr: 0.002633  loss: 3.0468 (3.0699)  class_acc: 0.5000 (0.5134)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9916 (2.5971)  time: 1.9409  data: 0.0006  max mem: 2905
Epoch: [131]  [400/625]  eta: 0:07:30  lr: 0.002626  min_lr: 0.002626  loss: 3.0858 (3.0833)  class_acc: 0.4922 (0.5101)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9900 (2.4006)  time: 1.8726  data: 0.0008  max mem: 2905
Epoch: [131]  [600/625]  eta: 0:00:50  lr: 0.002619  min_lr: 0.002619  loss: 3.0437 (3.0933)  class_acc: 0.5117 (0.5072)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8872 (2.4156)  time: 2.1263  data: 0.0007  max mem: 2905
Epoch: [131]  [624/625]  eta: 0:00:01  lr: 0.002618  min_lr: 0.002618  loss: 3.1339 (3.0944)  class_acc: 0.4922 (0.5068)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1859 (2.4276)  time: 0.6524  data: 0.0017  max mem: 2905
Epoch: [131] Total time: 0:20:22 (1.9561 s / it)
Averaged stats: lr: 0.002618  min_lr: 0.002618  loss: 3.1339 (3.0918)  class_acc: 0.4922 (0.5058)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1859 (2.4276)
Test:  [ 0/50]  eta: 0:10:49  loss: 3.8753 (3.8753)  acc1: 24.0000 (24.0000)  acc5: 53.6000 (53.6000)  time: 12.9968  data: 12.9594  max mem: 2905
Test:  [10/50]  eta: 0:01:25  loss: 3.7036 (3.6966)  acc1: 30.4000 (30.1091)  acc5: 56.8000 (55.4909)  time: 2.1387  data: 2.1133  max mem: 2905
Test:  [20/50]  eta: 0:00:49  loss: 3.7122 (3.8033)  acc1: 29.6000 (28.3048)  acc5: 54.4000 (53.7143)  time: 1.0947  data: 1.0727  max mem: 2905
Test:  [30/50]  eta: 0:00:29  loss: 3.8895 (3.8641)  acc1: 27.2000 (28.3355)  acc5: 50.4000 (52.2839)  time: 1.1186  data: 1.0983  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 3.8814 (3.8360)  acc1: 28.0000 (28.5659)  acc5: 50.4000 (52.4293)  time: 0.8823  data: 0.8622  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 3.7503 (3.8345)  acc1: 30.4000 (29.0240)  acc5: 52.0000 (52.4320)  time: 0.8088  data: 0.7881  max mem: 2905
Test: Total time: 0:00:53 (1.0619 s / it)
* Acc@1 29.670 Acc@5 52.654 loss 3.790
Accuracy of the model on the 50000 test images: 29.7%
Max accuracy: 50.54%
Epoch: [132]  [  0/625]  eta: 4:00:51  lr: 0.002618  min_lr: 0.002618  loss: 2.9817 (2.9817)  class_acc: 0.5312 (0.5312)  weight_decay: 0.0500 (0.0500)  time: 23.1219  data: 21.6330  max mem: 2905
Epoch: [132]  [200/625]  eta: 0:14:47  lr: 0.002612  min_lr: 0.002612  loss: 3.0838 (3.0779)  class_acc: 0.5000 (0.5057)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0978 (2.6972)  time: 1.8246  data: 1.2337  max mem: 2905
Epoch: [132]  [400/625]  eta: 0:07:35  lr: 0.002605  min_lr: 0.002605  loss: 3.1394 (3.0841)  class_acc: 0.4961 (0.5042)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0463 (2.5991)  time: 2.1092  data: 1.1952  max mem: 2905
Epoch: [132]  [600/625]  eta: 0:00:50  lr: 0.002598  min_lr: 0.002598  loss: 3.1011 (3.0896)  class_acc: 0.5078 (0.5043)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2059 (2.5178)  time: 1.9701  data: 1.7989  max mem: 2905
Epoch: [132]  [624/625]  eta: 0:00:01  lr: 0.002597  min_lr: 0.002597  loss: 3.1803 (3.0913)  class_acc: 0.4961 (0.5041)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2541 (inf)  time: 0.7571  data: 0.6053  max mem: 2905
Epoch: [132] Total time: 0:20:22 (1.9560 s / it)
Averaged stats: lr: 0.002597  min_lr: 0.002597  loss: 3.1803 (3.0907)  class_acc: 0.4961 (0.5059)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2541 (inf)
Test:  [ 0/50]  eta: 0:10:41  loss: 3.3563 (3.3563)  acc1: 34.4000 (34.4000)  acc5: 59.2000 (59.2000)  time: 12.8237  data: 12.7917  max mem: 2905
Test:  [10/50]  eta: 0:01:26  loss: 3.1198 (3.1489)  acc1: 38.4000 (37.9636)  acc5: 63.2000 (61.9636)  time: 2.1598  data: 2.1406  max mem: 2905
Test:  [20/50]  eta: 0:00:50  loss: 3.1198 (3.2114)  acc1: 38.4000 (37.2190)  acc5: 60.8000 (60.8762)  time: 1.1107  data: 1.0922  max mem: 2905
Test:  [30/50]  eta: 0:00:29  loss: 3.2023 (3.1852)  acc1: 36.8000 (37.2645)  acc5: 60.0000 (61.0065)  time: 1.0764  data: 1.0572  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 3.2023 (3.1928)  acc1: 38.4000 (37.5415)  acc5: 60.8000 (60.7024)  time: 0.7072  data: 0.6878  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 3.1572 (3.1698)  acc1: 38.4000 (37.6960)  acc5: 60.8000 (61.2480)  time: 0.6147  data: 0.5963  max mem: 2905
Test: Total time: 0:00:50 (1.0009 s / it)
* Acc@1 37.802 Acc@5 61.950 loss 3.152
Accuracy of the model on the 50000 test images: 37.8%
Max accuracy: 50.54%
Epoch: [133]  [  0/625]  eta: 3:47:41  lr: 0.002597  min_lr: 0.002597  loss: 2.8834 (2.8834)  class_acc: 0.5664 (0.5664)  weight_decay: 0.0500 (0.0500)  time: 21.8578  data: 18.2387  max mem: 2905
Epoch: [133]  [200/625]  eta: 0:13:49  lr: 0.002590  min_lr: 0.002590  loss: 3.0385 (3.0677)  class_acc: 0.5234 (0.5119)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3796 (2.3078)  time: 1.9566  data: 0.0010  max mem: 2905
Epoch: [133]  [400/625]  eta: 0:07:13  lr: 0.002583  min_lr: 0.002583  loss: 3.1106 (3.0745)  class_acc: 0.4922 (0.5093)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9767 (2.4886)  time: 1.9532  data: 0.0011  max mem: 2905
Epoch: [133]  [600/625]  eta: 0:00:49  lr: 0.002576  min_lr: 0.002576  loss: 3.0295 (3.0795)  class_acc: 0.5078 (0.5072)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1936 (2.5566)  time: 2.0735  data: 0.0018  max mem: 2905
Epoch: [133]  [624/625]  eta: 0:00:01  lr: 0.002576  min_lr: 0.002576  loss: 3.0714 (3.0792)  class_acc: 0.5156 (0.5072)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9940 (2.5608)  time: 0.7445  data: 0.0014  max mem: 2905
Epoch: [133] Total time: 0:19:57 (1.9164 s / it)
Averaged stats: lr: 0.002576  min_lr: 0.002576  loss: 3.0714 (3.0852)  class_acc: 0.5156 (0.5070)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9940 (2.5608)
Test:  [ 0/50]  eta: 0:10:22  loss: 2.9289 (2.9289)  acc1: 45.6000 (45.6000)  acc5: 64.0000 (64.0000)  time: 12.4445  data: 12.4172  max mem: 2905
Test:  [10/50]  eta: 0:01:29  loss: 2.9289 (2.9586)  acc1: 44.8000 (41.8909)  acc5: 64.8000 (65.2364)  time: 2.2397  data: 2.2186  max mem: 2905
Test:  [20/50]  eta: 0:00:54  loss: 3.1046 (3.0914)  acc1: 38.4000 (39.2762)  acc5: 63.2000 (62.8191)  time: 1.2882  data: 1.2673  max mem: 2905
Test:  [30/50]  eta: 0:00:32  loss: 3.1749 (3.1206)  acc1: 34.4000 (38.1419)  acc5: 60.8000 (62.2968)  time: 1.2659  data: 1.2458  max mem: 2905
Test:  [40/50]  eta: 0:00:13  loss: 3.1749 (3.1419)  acc1: 36.0000 (37.9512)  acc5: 60.8000 (61.9707)  time: 0.8894  data: 0.8710  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 3.1066 (3.1548)  acc1: 38.4000 (37.7920)  acc5: 62.4000 (62.2400)  time: 0.7562  data: 0.7356  max mem: 2905
Test: Total time: 0:00:57 (1.1431 s / it)
* Acc@1 37.562 Acc@5 62.364 loss 3.155
Accuracy of the model on the 50000 test images: 37.6%
Max accuracy: 50.54%
Epoch: [134]  [  0/625]  eta: 3:55:08  lr: 0.002576  min_lr: 0.002576  loss: 3.1418 (3.1418)  class_acc: 0.4922 (0.4922)  weight_decay: 0.0500 (0.0500)  time: 22.5737  data: 17.1659  max mem: 2905
Epoch: [134]  [200/625]  eta: 0:14:41  lr: 0.002569  min_lr: 0.002569  loss: 3.1136 (3.0636)  class_acc: 0.4961 (0.5117)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0655 (2.7874)  time: 2.0163  data: 0.0761  max mem: 2905
Epoch: [134]  [400/625]  eta: 0:07:31  lr: 0.002562  min_lr: 0.002562  loss: 3.0404 (3.0785)  class_acc: 0.5234 (0.5092)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0991 (2.5392)  time: 1.6611  data: 0.0552  max mem: 2905
Epoch: [134]  [600/625]  eta: 0:00:50  lr: 0.002555  min_lr: 0.002555  loss: 3.0568 (3.0820)  class_acc: 0.5195 (0.5086)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.8572 (2.5756)  time: 2.0468  data: 0.0266  max mem: 2905
Epoch: [134]  [624/625]  eta: 0:00:01  lr: 0.002554  min_lr: 0.002554  loss: 3.1062 (3.0840)  class_acc: 0.4961 (0.5082)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8271 (2.5438)  time: 0.7714  data: 0.0016  max mem: 2905
Epoch: [134] Total time: 0:20:30 (1.9695 s / it)
Averaged stats: lr: 0.002554  min_lr: 0.002554  loss: 3.1062 (3.0853)  class_acc: 0.4961 (0.5074)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8271 (2.5438)
Test:  [ 0/50]  eta: 0:10:34  loss: 2.8744 (2.8744)  acc1: 44.0000 (44.0000)  acc5: 64.0000 (64.0000)  time: 12.6816  data: 12.6517  max mem: 2905
Test:  [10/50]  eta: 0:01:27  loss: 2.7979 (2.7414)  acc1: 45.6000 (43.8545)  acc5: 66.4000 (68.4364)  time: 2.1864  data: 2.1664  max mem: 2905
Test:  [20/50]  eta: 0:00:51  loss: 2.8649 (2.8343)  acc1: 42.4000 (41.2952)  acc5: 66.4000 (67.8095)  time: 1.1763  data: 1.1566  max mem: 2905
Test:  [30/50]  eta: 0:00:30  loss: 2.8933 (2.8228)  acc1: 41.6000 (41.6000)  acc5: 65.6000 (67.4839)  time: 1.1369  data: 1.1173  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 2.9831 (2.8686)  acc1: 40.0000 (41.1512)  acc5: 62.4000 (66.5561)  time: 0.6994  data: 0.6796  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.8160 (2.8491)  acc1: 40.0000 (41.2000)  acc5: 65.6000 (67.1040)  time: 0.5741  data: 0.5543  max mem: 2905
Test: Total time: 0:00:51 (1.0231 s / it)
* Acc@1 42.118 Acc@5 67.348 loss 2.825
Accuracy of the model on the 50000 test images: 42.1%
Max accuracy: 50.54%
Epoch: [135]  [  0/625]  eta: 3:30:14  lr: 0.002554  min_lr: 0.002554  loss: 2.9742 (2.9742)  class_acc: 0.5469 (0.5469)  weight_decay: 0.0500 (0.0500)  time: 20.1824  data: 18.0960  max mem: 2905
Epoch: [135]  [200/625]  eta: 0:14:28  lr: 0.002547  min_lr: 0.002547  loss: 3.0613 (3.0633)  class_acc: 0.5039 (0.5089)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.6176 (2.4505)  time: 1.8430  data: 0.0007  max mem: 2905
Epoch: [135]  [400/625]  eta: 0:07:28  lr: 0.002540  min_lr: 0.002540  loss: 3.0230 (3.0801)  class_acc: 0.5195 (0.5067)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.6732 (2.5765)  time: 2.0788  data: 0.0113  max mem: 2905
Epoch: [135]  [600/625]  eta: 0:00:49  lr: 0.002533  min_lr: 0.002533  loss: 3.1288 (3.0886)  class_acc: 0.4961 (0.5053)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2097 (2.5779)  time: 1.9825  data: 0.0107  max mem: 2905
Epoch: [135]  [624/625]  eta: 0:00:01  lr: 0.002533  min_lr: 0.002533  loss: 3.1344 (3.0895)  class_acc: 0.5039 (0.5050)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0216 (2.5654)  time: 0.8351  data: 0.0013  max mem: 2905
Epoch: [135] Total time: 0:20:04 (1.9264 s / it)
Averaged stats: lr: 0.002533  min_lr: 0.002533  loss: 3.1344 (3.0839)  class_acc: 0.5039 (0.5075)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0216 (2.5654)
Test:  [ 0/50]  eta: 0:09:08  loss: 2.6195 (2.6195)  acc1: 48.8000 (48.8000)  acc5: 72.8000 (72.8000)  time: 10.9628  data: 10.9350  max mem: 2905
Test:  [10/50]  eta: 0:01:18  loss: 2.6195 (2.5478)  acc1: 48.8000 (48.0727)  acc5: 72.8000 (72.0000)  time: 1.9565  data: 1.9366  max mem: 2905
Test:  [20/50]  eta: 0:00:47  loss: 2.7389 (2.7370)  acc1: 42.4000 (44.4190)  acc5: 67.2000 (68.8381)  time: 1.0976  data: 1.0786  max mem: 2905
Test:  [30/50]  eta: 0:00:27  loss: 2.6896 (2.7006)  acc1: 41.6000 (44.6452)  acc5: 67.2000 (69.7290)  time: 1.0168  data: 0.9965  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 2.6447 (2.6930)  acc1: 45.6000 (45.0927)  acc5: 68.8000 (69.7366)  time: 0.6306  data: 0.6095  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 2.6632 (2.6966)  acc1: 44.0000 (44.6560)  acc5: 68.8000 (69.6000)  time: 0.5742  data: 0.5546  max mem: 2905
Test: Total time: 0:00:46 (0.9361 s / it)
* Acc@1 44.968 Acc@5 70.110 loss 2.668
Accuracy of the model on the 50000 test images: 45.0%
Max accuracy: 50.54%
Epoch: [136]  [  0/625]  eta: 3:35:22  lr: 0.002532  min_lr: 0.002532  loss: 2.8887 (2.8887)  class_acc: 0.5273 (0.5273)  weight_decay: 0.0500 (0.0500)  time: 20.6767  data: 17.5166  max mem: 2905
Epoch: [136]  [200/625]  eta: 0:13:44  lr: 0.002526  min_lr: 0.002526  loss: 3.0730 (3.0530)  class_acc: 0.5039 (0.5150)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8825 (2.5561)  time: 1.7947  data: 0.0351  max mem: 2905
Epoch: [136]  [400/625]  eta: 0:07:11  lr: 0.002519  min_lr: 0.002519  loss: 3.0894 (3.0723)  class_acc: 0.5117 (0.5099)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1014 (2.5203)  time: 1.9860  data: 0.0515  max mem: 2905
Epoch: [136]  [600/625]  eta: 0:00:47  lr: 0.002512  min_lr: 0.002512  loss: 3.0741 (3.0770)  class_acc: 0.5078 (0.5089)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2991 (2.4746)  time: 1.7775  data: 0.2486  max mem: 2905
Epoch: [136]  [624/625]  eta: 0:00:01  lr: 0.002511  min_lr: 0.002511  loss: 3.0776 (3.0774)  class_acc: 0.5117 (0.5090)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5085 (2.4902)  time: 0.7786  data: 0.3327  max mem: 2905
Epoch: [136] Total time: 0:19:24 (1.8627 s / it)
Averaged stats: lr: 0.002511  min_lr: 0.002511  loss: 3.0776 (3.0764)  class_acc: 0.5117 (0.5091)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5085 (2.4902)
Test:  [ 0/50]  eta: 0:10:56  loss: 2.8796 (2.8796)  acc1: 37.6000 (37.6000)  acc5: 64.8000 (64.8000)  time: 13.1252  data: 13.0958  max mem: 2905
Test:  [10/50]  eta: 0:01:21  loss: 2.6973 (2.6960)  acc1: 43.2000 (43.5636)  acc5: 67.2000 (69.0909)  time: 2.0297  data: 2.0104  max mem: 2905
Test:  [20/50]  eta: 0:00:46  loss: 2.9880 (2.9792)  acc1: 37.6000 (39.3524)  acc5: 64.0000 (65.4095)  time: 0.9670  data: 0.9483  max mem: 2905
Test:  [30/50]  eta: 0:00:27  loss: 3.1363 (3.0198)  acc1: 35.2000 (39.1742)  acc5: 60.8000 (64.7226)  time: 0.9917  data: 0.9726  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 3.0299 (3.0329)  acc1: 35.2000 (38.8293)  acc5: 64.0000 (64.1171)  time: 0.8095  data: 0.7905  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 3.1343 (3.1084)  acc1: 34.4000 (37.7760)  acc5: 61.6000 (63.1680)  time: 0.8178  data: 0.7969  max mem: 2905
Test: Total time: 0:00:51 (1.0226 s / it)
* Acc@1 38.236 Acc@5 63.560 loss 3.092
Accuracy of the model on the 50000 test images: 38.2%
Max accuracy: 50.54%
Epoch: [137]  [  0/625]  eta: 3:33:32  lr: 0.002511  min_lr: 0.002511  loss: 3.1157 (3.1157)  class_acc: 0.4766 (0.4766)  weight_decay: 0.0500 (0.0500)  time: 20.5002  data: 17.0115  max mem: 2905
Epoch: [137]  [200/625]  eta: 0:13:44  lr: 0.002504  min_lr: 0.002504  loss: 3.0574 (3.0515)  class_acc: 0.5117 (0.5149)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4387 (2.4823)  time: 1.8340  data: 0.0006  max mem: 2905
Epoch: [137]  [400/625]  eta: 0:07:15  lr: 0.002497  min_lr: 0.002497  loss: 3.1327 (3.0634)  class_acc: 0.4922 (0.5108)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9285 (2.4755)  time: 1.9271  data: 0.0007  max mem: 2905
Epoch: [137]  [600/625]  eta: 0:00:48  lr: 0.002490  min_lr: 0.002490  loss: 3.0988 (3.0693)  class_acc: 0.5039 (0.5111)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7438 (2.5150)  time: 1.9791  data: 0.0008  max mem: 2905
Epoch: [137]  [624/625]  eta: 0:00:01  lr: 0.002489  min_lr: 0.002489  loss: 3.1064 (3.0705)  class_acc: 0.5078 (0.5111)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5806 (2.5282)  time: 0.8149  data: 0.1237  max mem: 2905
Epoch: [137] Total time: 0:19:46 (1.8978 s / it)
Averaged stats: lr: 0.002489  min_lr: 0.002489  loss: 3.1064 (3.0754)  class_acc: 0.5078 (0.5093)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5806 (2.5282)
Test:  [ 0/50]  eta: 0:10:17  loss: 3.4693 (3.4693)  acc1: 35.2000 (35.2000)  acc5: 60.8000 (60.8000)  time: 12.3499  data: 12.3215  max mem: 2905
Test:  [10/50]  eta: 0:01:29  loss: 3.0785 (2.9895)  acc1: 38.4000 (40.4364)  acc5: 63.2000 (65.5273)  time: 2.2368  data: 2.2153  max mem: 2905
Test:  [20/50]  eta: 0:00:52  loss: 3.0792 (3.1107)  acc1: 36.8000 (38.2476)  acc5: 62.4000 (63.4286)  time: 1.2124  data: 1.1916  max mem: 2905
Test:  [30/50]  eta: 0:00:29  loss: 3.1595 (3.1155)  acc1: 36.8000 (38.6581)  acc5: 60.0000 (63.2774)  time: 1.0249  data: 1.0048  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 3.1201 (3.1650)  acc1: 37.6000 (37.8927)  acc5: 60.0000 (62.4781)  time: 0.6029  data: 0.5837  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 2.9527 (3.1231)  acc1: 38.4000 (38.6880)  acc5: 64.0000 (63.2320)  time: 0.5648  data: 0.5465  max mem: 2905
Test: Total time: 0:00:49 (0.9882 s / it)
* Acc@1 38.972 Acc@5 63.514 loss 3.087
Accuracy of the model on the 50000 test images: 39.0%
Max accuracy: 50.54%
Epoch: [138]  [  0/625]  eta: 4:30:32  lr: 0.002489  min_lr: 0.002489  loss: 2.9704 (2.9704)  class_acc: 0.5352 (0.5352)  weight_decay: 0.0500 (0.0500)  time: 25.9724  data: 19.9072  max mem: 2905
Epoch: [138]  [200/625]  eta: 0:14:23  lr: 0.002482  min_lr: 0.002482  loss: 3.0487 (3.0520)  class_acc: 0.5117 (0.5141)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9061 (2.4300)  time: 1.9446  data: 0.0229  max mem: 2905
Epoch: [138]  [400/625]  eta: 0:07:25  lr: 0.002475  min_lr: 0.002475  loss: 3.1225 (3.0603)  class_acc: 0.4961 (0.5122)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1429 (2.5400)  time: 1.9835  data: 0.0110  max mem: 2905
Epoch: [138]  [600/625]  eta: 0:00:49  lr: 0.002468  min_lr: 0.002468  loss: 3.0555 (3.0663)  class_acc: 0.5000 (0.5112)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7596 (2.5965)  time: 1.8843  data: 0.0007  max mem: 2905
Epoch: [138]  [624/625]  eta: 0:00:01  lr: 0.002467  min_lr: 0.002467  loss: 3.1084 (3.0688)  class_acc: 0.5039 (0.5108)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4343 (2.5907)  time: 0.9180  data: 0.0148  max mem: 2905
Epoch: [138] Total time: 0:20:07 (1.9324 s / it)
Averaged stats: lr: 0.002467  min_lr: 0.002467  loss: 3.1084 (3.0770)  class_acc: 0.5039 (0.5087)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4343 (2.5907)
Test:  [ 0/50]  eta: 0:09:15  loss: 3.5574 (3.5574)  acc1: 30.4000 (30.4000)  acc5: 53.6000 (53.6000)  time: 11.1032  data: 11.0800  max mem: 2905
Test:  [10/50]  eta: 0:01:13  loss: 3.3169 (3.3655)  acc1: 36.0000 (35.5636)  acc5: 57.6000 (57.8182)  time: 1.8332  data: 1.8135  max mem: 2905
Test:  [20/50]  eta: 0:00:43  loss: 3.4039 (3.4427)  acc1: 32.8000 (34.0952)  acc5: 57.6000 (57.1810)  time: 0.9577  data: 0.9368  max mem: 2905
Test:  [30/50]  eta: 0:00:25  loss: 3.5441 (3.5067)  acc1: 30.4000 (33.2387)  acc5: 56.0000 (57.0065)  time: 0.9812  data: 0.9603  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 3.7328 (3.5416)  acc1: 30.4000 (32.7220)  acc5: 56.8000 (56.6829)  time: 0.8310  data: 0.8121  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 3.5819 (3.5300)  acc1: 31.2000 (33.0240)  acc5: 56.8000 (56.9440)  time: 0.7929  data: 0.7740  max mem: 2905
Test: Total time: 0:00:50 (1.0049 s / it)
* Acc@1 33.694 Acc@5 57.450 loss 3.479
Accuracy of the model on the 50000 test images: 33.7%
Max accuracy: 50.54%
Epoch: [139]  [  0/625]  eta: 4:15:13  lr: 0.002467  min_lr: 0.002467  loss: 2.9521 (2.9521)  class_acc: 0.5195 (0.5195)  weight_decay: 0.0500 (0.0500)  time: 24.5017  data: 18.5787  max mem: 2905
Epoch: [139]  [200/625]  eta: 0:14:36  lr: 0.002460  min_lr: 0.002460  loss: 3.1134 (3.0511)  class_acc: 0.5117 (0.5145)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0421 (2.3328)  time: 1.8900  data: 0.0009  max mem: 2905
Epoch: [139]  [400/625]  eta: 0:07:29  lr: 0.002453  min_lr: 0.002453  loss: 3.0990 (3.0677)  class_acc: 0.4922 (0.5111)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5744 (2.4046)  time: 2.0204  data: 0.0827  max mem: 2905
Epoch: [139]  [600/625]  eta: 0:00:50  lr: 0.002446  min_lr: 0.002446  loss: 3.0533 (3.0758)  class_acc: 0.5000 (0.5089)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6337 (inf)  time: 1.9939  data: 0.0009  max mem: 2905
Epoch: [139]  [624/625]  eta: 0:00:01  lr: 0.002446  min_lr: 0.002446  loss: 3.0955 (3.0768)  class_acc: 0.5000 (0.5088)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3258 (inf)  time: 0.7435  data: 0.0015  max mem: 2905
Epoch: [139] Total time: 0:20:22 (1.9560 s / it)
Averaged stats: lr: 0.002446  min_lr: 0.002446  loss: 3.0955 (3.0721)  class_acc: 0.5000 (0.5104)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3258 (inf)
Test:  [ 0/50]  eta: 0:09:43  loss: 2.7553 (2.7553)  acc1: 39.2000 (39.2000)  acc5: 63.2000 (63.2000)  time: 11.6684  data: 11.6374  max mem: 2905
Test:  [10/50]  eta: 0:01:25  loss: 3.2310 (3.2752)  acc1: 38.4000 (37.7455)  acc5: 60.8000 (60.0727)  time: 2.1353  data: 2.1149  max mem: 2905
Test:  [20/50]  eta: 0:00:51  loss: 3.4882 (3.4361)  acc1: 33.6000 (34.6667)  acc5: 56.8000 (57.9429)  time: 1.2273  data: 1.2081  max mem: 2905
Test:  [30/50]  eta: 0:00:29  loss: 3.7071 (3.4906)  acc1: 31.2000 (34.0129)  acc5: 55.2000 (57.1871)  time: 1.1301  data: 1.1094  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 3.6503 (3.5432)  acc1: 32.0000 (33.6195)  acc5: 54.4000 (56.2341)  time: 0.6680  data: 0.6462  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 3.5699 (3.5565)  acc1: 32.0000 (33.1040)  acc5: 55.2000 (56.2560)  time: 0.6150  data: 0.5931  max mem: 2905
Test: Total time: 0:00:50 (1.0074 s / it)
* Acc@1 33.010 Acc@5 56.616 loss 3.518
Accuracy of the model on the 50000 test images: 33.0%
Max accuracy: 50.54%
Epoch: [140]  [  0/625]  eta: 4:04:05  lr: 0.002445  min_lr: 0.002445  loss: 3.1129 (3.1129)  class_acc: 0.4805 (0.4805)  weight_decay: 0.0500 (0.0500)  time: 23.4328  data: 20.6775  max mem: 2905
Epoch: [140]  [200/625]  eta: 0:14:33  lr: 0.002438  min_lr: 0.002438  loss: 3.1190 (3.0637)  class_acc: 0.5039 (0.5121)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1149 (2.2286)  time: 1.8908  data: 0.0006  max mem: 2905
Epoch: [140]  [400/625]  eta: 0:07:37  lr: 0.002431  min_lr: 0.002431  loss: 3.0820 (3.0676)  class_acc: 0.5000 (0.5100)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.1054 (2.4204)  time: 1.9468  data: 0.0009  max mem: 2905
Epoch: [140]  [600/625]  eta: 0:00:50  lr: 0.002424  min_lr: 0.002424  loss: 3.1091 (3.0742)  class_acc: 0.4883 (0.5089)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8624 (2.4511)  time: 1.9886  data: 0.0008  max mem: 2905
Epoch: [140]  [624/625]  eta: 0:00:01  lr: 0.002424  min_lr: 0.002424  loss: 3.0889 (3.0742)  class_acc: 0.5078 (0.5088)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0453 (2.4640)  time: 0.8653  data: 0.0014  max mem: 2905
Epoch: [140] Total time: 0:20:34 (1.9744 s / it)
Averaged stats: lr: 0.002424  min_lr: 0.002424  loss: 3.0889 (3.0672)  class_acc: 0.5078 (0.5112)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0453 (2.4640)
Test:  [ 0/50]  eta: 0:09:31  loss: 2.6982 (2.6982)  acc1: 43.2000 (43.2000)  acc5: 71.2000 (71.2000)  time: 11.4348  data: 11.4081  max mem: 2905
Test:  [10/50]  eta: 0:01:19  loss: 2.8186 (2.8873)  acc1: 42.4000 (42.3273)  acc5: 64.8000 (65.6727)  time: 1.9786  data: 1.9575  max mem: 2905
Test:  [20/50]  eta: 0:00:47  loss: 3.0463 (3.0889)  acc1: 37.6000 (38.4000)  acc5: 61.6000 (63.4667)  time: 1.0915  data: 1.0718  max mem: 2905
Test:  [30/50]  eta: 0:00:28  loss: 3.2969 (3.1485)  acc1: 34.4000 (37.1355)  acc5: 60.8000 (62.3484)  time: 1.0894  data: 1.0703  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 3.4500 (3.2531)  acc1: 30.4000 (35.3366)  acc5: 57.6000 (60.7610)  time: 0.8052  data: 0.7859  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 3.3255 (3.2708)  acc1: 31.2000 (34.9920)  acc5: 58.4000 (60.6080)  time: 0.7198  data: 0.7010  max mem: 2905
Test: Total time: 0:00:50 (1.0103 s / it)
* Acc@1 35.424 Acc@5 60.910 loss 3.244
Accuracy of the model on the 50000 test images: 35.4%
Max accuracy: 50.54%
Epoch: [141]  [  0/625]  eta: 3:31:49  lr: 0.002424  min_lr: 0.002424  loss: 2.8883 (2.8883)  class_acc: 0.5664 (0.5664)  weight_decay: 0.0500 (0.0500)  time: 20.3349  data: 18.4900  max mem: 2905
Epoch: [141]  [200/625]  eta: 0:14:15  lr: 0.002417  min_lr: 0.002417  loss: 3.0460 (3.0330)  class_acc: 0.5156 (0.5191)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8336 (2.5539)  time: 1.7892  data: 0.1157  max mem: 2905
Epoch: [141]  [400/625]  eta: 0:07:23  lr: 0.002409  min_lr: 0.002409  loss: 3.0470 (3.0547)  class_acc: 0.5000 (0.5134)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1578 (2.4908)  time: 1.8894  data: 0.0324  max mem: 2905
Epoch: [141]  [600/625]  eta: 0:00:48  lr: 0.002402  min_lr: 0.002402  loss: 3.0643 (3.0666)  class_acc: 0.5039 (0.5114)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0136 (2.5000)  time: 1.9136  data: 0.0007  max mem: 2905
Epoch: [141]  [624/625]  eta: 0:00:01  lr: 0.002402  min_lr: 0.002402  loss: 3.0441 (3.0652)  class_acc: 0.5195 (0.5117)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3080 (2.5106)  time: 0.8594  data: 0.0013  max mem: 2905
Epoch: [141] Total time: 0:19:55 (1.9133 s / it)
Averaged stats: lr: 0.002402  min_lr: 0.002402  loss: 3.0441 (3.0653)  class_acc: 0.5195 (0.5119)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3080 (2.5106)
Test:  [ 0/50]  eta: 0:09:44  loss: 2.8425 (2.8425)  acc1: 40.8000 (40.8000)  acc5: 65.6000 (65.6000)  time: 11.6981  data: 11.6530  max mem: 2905
Test:  [10/50]  eta: 0:01:21  loss: 2.7544 (2.7566)  acc1: 42.4000 (43.4182)  acc5: 66.4000 (68.2909)  time: 2.0483  data: 2.0258  max mem: 2905
Test:  [20/50]  eta: 0:00:48  loss: 2.9078 (2.9245)  acc1: 39.2000 (40.0381)  acc5: 68.0000 (66.5143)  time: 1.1187  data: 1.0991  max mem: 2905
Test:  [30/50]  eta: 0:00:28  loss: 2.9040 (2.9074)  acc1: 39.2000 (40.6710)  acc5: 66.4000 (66.1677)  time: 1.1138  data: 1.0937  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 2.8411 (2.9284)  acc1: 40.0000 (40.5659)  acc5: 62.4000 (65.8146)  time: 0.7600  data: 0.7403  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 2.9588 (2.9424)  acc1: 38.4000 (40.4960)  acc5: 64.0000 (65.7280)  time: 0.7342  data: 0.7158  max mem: 2905
Test: Total time: 0:00:49 (0.9977 s / it)
* Acc@1 40.812 Acc@5 66.092 loss 2.915
Accuracy of the model on the 50000 test images: 40.8%
Max accuracy: 50.54%
Epoch: [142]  [  0/625]  eta: 3:39:07  lr: 0.002402  min_lr: 0.002402  loss: 3.0267 (3.0267)  class_acc: 0.5039 (0.5039)  weight_decay: 0.0500 (0.0500)  time: 21.0363  data: 18.1838  max mem: 2905
Epoch: [142]  [200/625]  eta: 0:13:46  lr: 0.002395  min_lr: 0.002395  loss: 3.0032 (3.0524)  class_acc: 0.5234 (0.5137)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4480 (2.6205)  time: 1.8363  data: 0.0007  max mem: 2905
Epoch: [142]  [400/625]  eta: 0:07:10  lr: 0.002387  min_lr: 0.002387  loss: 3.0138 (3.0590)  class_acc: 0.5234 (0.5129)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5212 (2.5400)  time: 1.8549  data: 0.0008  max mem: 2905
Epoch: [142]  [600/625]  eta: 0:00:48  lr: 0.002380  min_lr: 0.002380  loss: 3.0526 (3.0600)  class_acc: 0.5117 (0.5132)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4879 (2.5871)  time: 2.0143  data: 0.0464  max mem: 2905
Epoch: [142]  [624/625]  eta: 0:00:01  lr: 0.002380  min_lr: 0.002380  loss: 3.0838 (3.0604)  class_acc: 0.5117 (0.5131)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5098 (2.5811)  time: 0.8553  data: 0.0324  max mem: 2905
Epoch: [142] Total time: 0:19:36 (1.8829 s / it)
Averaged stats: lr: 0.002380  min_lr: 0.002380  loss: 3.0838 (3.0624)  class_acc: 0.5117 (0.5125)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5098 (2.5811)
Test:  [ 0/50]  eta: 0:10:35  loss: 2.5529 (2.5529)  acc1: 45.6000 (45.6000)  acc5: 76.0000 (76.0000)  time: 12.7139  data: 12.6900  max mem: 2905
Test:  [10/50]  eta: 0:01:29  loss: 2.7295 (2.7750)  acc1: 45.6000 (44.1455)  acc5: 67.2000 (67.7818)  time: 2.2323  data: 2.2109  max mem: 2905
Test:  [20/50]  eta: 0:00:53  loss: 2.9010 (2.9018)  acc1: 40.8000 (40.8381)  acc5: 65.6000 (66.6667)  time: 1.2302  data: 1.2102  max mem: 2905
Test:  [30/50]  eta: 0:00:32  loss: 2.9971 (2.9226)  acc1: 38.4000 (40.7484)  acc5: 64.0000 (66.4258)  time: 1.2898  data: 1.2706  max mem: 2905
Test:  [40/50]  eta: 0:00:13  loss: 2.9971 (2.9448)  acc1: 37.6000 (40.3707)  acc5: 64.0000 (65.9317)  time: 0.8946  data: 0.8735  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.9262 (2.9244)  acc1: 38.4000 (40.7840)  acc5: 64.8000 (66.1280)  time: 0.8447  data: 0.8237  max mem: 2905
Test: Total time: 0:00:55 (1.1161 s / it)
* Acc@1 41.306 Acc@5 66.844 loss 2.901
Accuracy of the model on the 50000 test images: 41.3%
Max accuracy: 50.54%
Epoch: [143]  [  0/625]  eta: 3:47:01  lr: 0.002380  min_lr: 0.002380  loss: 3.0527 (3.0527)  class_acc: 0.5195 (0.5195)  weight_decay: 0.0500 (0.0500)  time: 21.7950  data: 19.6368  max mem: 2905
Epoch: [143]  [200/625]  eta: 0:14:21  lr: 0.002373  min_lr: 0.002373  loss: 3.0497 (3.0543)  class_acc: 0.5039 (0.5142)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0981 (2.3357)  time: 1.7320  data: 0.0229  max mem: 2905
Epoch: [143]  [400/625]  eta: 0:07:23  lr: 0.002365  min_lr: 0.002365  loss: 3.0871 (3.0539)  class_acc: 0.5000 (0.5146)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3299 (2.3092)  time: 1.9094  data: 0.0281  max mem: 2905
Epoch: [143]  [600/625]  eta: 0:00:48  lr: 0.002358  min_lr: 0.002358  loss: 3.0476 (3.0592)  class_acc: 0.5078 (0.5131)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7280 (2.4605)  time: 2.0905  data: 0.0351  max mem: 2905
Epoch: [143]  [624/625]  eta: 0:00:01  lr: 0.002358  min_lr: 0.002358  loss: 3.0777 (3.0605)  class_acc: 0.5039 (0.5128)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7995 (2.4555)  time: 0.8065  data: 0.0190  max mem: 2905
Epoch: [143] Total time: 0:19:56 (1.9137 s / it)
Averaged stats: lr: 0.002358  min_lr: 0.002358  loss: 3.0777 (3.0604)  class_acc: 0.5039 (0.5126)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7995 (2.4555)
Test:  [ 0/50]  eta: 0:10:49  loss: 2.8117 (2.8117)  acc1: 39.2000 (39.2000)  acc5: 68.8000 (68.8000)  time: 12.9843  data: 12.9552  max mem: 2905
Test:  [10/50]  eta: 0:01:21  loss: 2.8304 (2.8467)  acc1: 41.6000 (41.5273)  acc5: 67.2000 (67.1273)  time: 2.0420  data: 2.0209  max mem: 2905
Test:  [20/50]  eta: 0:00:46  loss: 3.0952 (3.0441)  acc1: 38.4000 (38.8190)  acc5: 63.2000 (64.3429)  time: 0.9826  data: 0.9632  max mem: 2905
Test:  [30/50]  eta: 0:00:26  loss: 3.1793 (3.0280)  acc1: 37.6000 (39.4323)  acc5: 63.2000 (64.7226)  time: 0.9622  data: 0.9436  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 3.1292 (3.0684)  acc1: 37.6000 (38.7122)  acc5: 63.2000 (63.9610)  time: 0.6577  data: 0.6391  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 3.1208 (3.0727)  acc1: 37.6000 (38.7680)  acc5: 63.2000 (63.9200)  time: 0.5875  data: 0.5691  max mem: 2905
Test: Total time: 0:00:47 (0.9456 s / it)
* Acc@1 39.052 Acc@5 64.066 loss 3.030
Accuracy of the model on the 50000 test images: 39.1%
Max accuracy: 50.54%
Epoch: [144]  [  0/625]  eta: 3:31:47  lr: 0.002358  min_lr: 0.002358  loss: 3.0349 (3.0349)  class_acc: 0.5352 (0.5352)  weight_decay: 0.0500 (0.0500)  time: 20.3321  data: 18.2587  max mem: 2905
Epoch: [144]  [200/625]  eta: 0:14:12  lr: 0.002350  min_lr: 0.002350  loss: 2.9920 (3.0409)  class_acc: 0.5156 (0.5182)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1870 (2.4454)  time: 1.8720  data: 0.0057  max mem: 2905
Epoch: [144]  [400/625]  eta: 0:07:23  lr: 0.002343  min_lr: 0.002343  loss: 3.0716 (3.0474)  class_acc: 0.5156 (0.5145)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0637 (2.5605)  time: 2.0267  data: 0.0659  max mem: 2905
Epoch: [144]  [600/625]  eta: 0:00:49  lr: 0.002336  min_lr: 0.002336  loss: 3.0663 (3.0522)  class_acc: 0.5195 (0.5142)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7558 (2.5100)  time: 2.0571  data: 0.0037  max mem: 2905
Epoch: [144]  [624/625]  eta: 0:00:01  lr: 0.002335  min_lr: 0.002335  loss: 3.1179 (3.0541)  class_acc: 0.4922 (0.5138)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7494 (2.5205)  time: 0.9005  data: 0.0095  max mem: 2905
Epoch: [144] Total time: 0:20:13 (1.9423 s / it)
Averaged stats: lr: 0.002335  min_lr: 0.002335  loss: 3.1179 (3.0561)  class_acc: 0.4922 (0.5136)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7494 (2.5205)
Test:  [ 0/50]  eta: 0:09:26  loss: 3.0690 (3.0690)  acc1: 35.2000 (35.2000)  acc5: 64.0000 (64.0000)  time: 11.3384  data: 11.3129  max mem: 2905
Test:  [10/50]  eta: 0:01:13  loss: 3.0585 (2.9509)  acc1: 40.0000 (40.3636)  acc5: 64.8000 (64.5818)  time: 1.8457  data: 1.8273  max mem: 2905
Test:  [20/50]  eta: 0:00:43  loss: 3.0100 (3.0383)  acc1: 39.2000 (38.2857)  acc5: 64.8000 (63.1238)  time: 0.9509  data: 0.9328  max mem: 2905
Test:  [30/50]  eta: 0:00:26  loss: 3.0654 (3.0544)  acc1: 36.8000 (38.6065)  acc5: 63.2000 (63.2516)  time: 1.0023  data: 0.9832  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 3.0751 (3.0647)  acc1: 39.2000 (38.4000)  acc5: 64.0000 (63.2585)  time: 1.0781  data: 1.0594  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 3.0188 (3.0677)  acc1: 36.0000 (38.1760)  acc5: 63.2000 (63.3280)  time: 0.8005  data: 0.7826  max mem: 2905
Test: Total time: 0:00:56 (1.1317 s / it)
* Acc@1 39.424 Acc@5 64.080 loss 3.020
Accuracy of the model on the 50000 test images: 39.4%
Max accuracy: 50.54%
Epoch: [145]  [  0/625]  eta: 3:36:02  lr: 0.002335  min_lr: 0.002335  loss: 2.8599 (2.8599)  class_acc: 0.5430 (0.5430)  weight_decay: 0.0500 (0.0500)  time: 20.7404  data: 17.3197  max mem: 2905
Epoch: [145]  [200/625]  eta: 0:14:27  lr: 0.002328  min_lr: 0.002328  loss: 2.9824 (3.0238)  class_acc: 0.5234 (0.5212)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1198 (2.4618)  time: 1.9230  data: 0.0068  max mem: 2905
Epoch: [145]  [400/625]  eta: 0:07:31  lr: 0.002321  min_lr: 0.002321  loss: 3.0434 (3.0445)  class_acc: 0.5078 (0.5168)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5601 (2.5921)  time: 2.1497  data: 0.0007  max mem: 2905
Epoch: [145]  [600/625]  eta: 0:00:50  lr: 0.002314  min_lr: 0.002314  loss: 3.0194 (3.0510)  class_acc: 0.5195 (0.5156)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1357 (2.5381)  time: 2.1454  data: 0.0150  max mem: 2905
Epoch: [145]  [624/625]  eta: 0:00:01  lr: 0.002313  min_lr: 0.002313  loss: 3.0428 (3.0504)  class_acc: 0.5117 (0.5156)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3087 (2.5419)  time: 0.6577  data: 0.0018  max mem: 2905
Epoch: [145] Total time: 0:20:18 (1.9490 s / it)
Averaged stats: lr: 0.002313  min_lr: 0.002313  loss: 3.0428 (3.0521)  class_acc: 0.5117 (0.5148)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3087 (2.5419)
Test:  [ 0/50]  eta: 0:10:06  loss: 3.0126 (3.0126)  acc1: 39.2000 (39.2000)  acc5: 64.8000 (64.8000)  time: 12.1301  data: 12.1020  max mem: 2905
Test:  [10/50]  eta: 0:01:10  loss: 3.0126 (3.0161)  acc1: 39.2000 (40.4364)  acc5: 64.8000 (64.2909)  time: 1.7690  data: 1.7470  max mem: 2905
Test:  [20/50]  eta: 0:00:38  loss: 3.2252 (3.2251)  acc1: 33.6000 (36.5333)  acc5: 60.0000 (60.8762)  time: 0.7394  data: 0.7179  max mem: 2905
Test:  [30/50]  eta: 0:00:23  loss: 3.2189 (3.1635)  acc1: 33.6000 (37.1613)  acc5: 60.8000 (61.9871)  time: 0.8651  data: 0.8452  max mem: 2905
Test:  [40/50]  eta: 0:00:10  loss: 3.1708 (3.2238)  acc1: 35.2000 (36.5463)  acc5: 61.6000 (61.1707)  time: 0.8513  data: 0.8336  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 3.2854 (3.2300)  acc1: 33.6000 (36.6080)  acc5: 60.0000 (61.1360)  time: 0.5136  data: 0.4959  max mem: 2905
Test: Total time: 0:00:47 (0.9433 s / it)
* Acc@1 37.644 Acc@5 61.944 loss 3.189
Accuracy of the model on the 50000 test images: 37.6%
Max accuracy: 50.54%
Epoch: [146]  [  0/625]  eta: 3:55:02  lr: 0.002313  min_lr: 0.002313  loss: 3.0128 (3.0128)  class_acc: 0.5195 (0.5195)  weight_decay: 0.0500 (0.0500)  time: 22.5640  data: 21.0014  max mem: 2905
Epoch: [146]  [200/625]  eta: 0:14:47  lr: 0.002306  min_lr: 0.002306  loss: 3.0039 (3.0538)  class_acc: 0.5117 (0.5141)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9791 (2.6369)  time: 1.8616  data: 0.1726  max mem: 2905
Epoch: [146]  [400/625]  eta: 0:07:33  lr: 0.002299  min_lr: 0.002299  loss: 3.0296 (3.0546)  class_acc: 0.5117 (0.5150)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5003 (inf)  time: 1.9097  data: 0.0032  max mem: 2905
Epoch: [146]  [600/625]  eta: 0:00:50  lr: 0.002292  min_lr: 0.002292  loss: 3.0597 (3.0568)  class_acc: 0.5039 (0.5142)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8707 (inf)  time: 2.0692  data: 0.0098  max mem: 2905
Epoch: [146]  [624/625]  eta: 0:00:01  lr: 0.002291  min_lr: 0.002291  loss: 3.0740 (3.0579)  class_acc: 0.5039 (0.5138)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.6800 (inf)  time: 0.7919  data: 0.0015  max mem: 2905
Epoch: [146] Total time: 0:20:29 (1.9674 s / it)
Averaged stats: lr: 0.002291  min_lr: 0.002291  loss: 3.0740 (3.0554)  class_acc: 0.5039 (0.5138)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.6800 (inf)
Test:  [ 0/50]  eta: 0:09:44  loss: 3.3146 (3.3146)  acc1: 37.6000 (37.6000)  acc5: 60.0000 (60.0000)  time: 11.6938  data: 11.6670  max mem: 2905
Test:  [10/50]  eta: 0:01:17  loss: 3.5712 (3.6294)  acc1: 32.8000 (32.9455)  acc5: 53.6000 (54.9091)  time: 1.9435  data: 1.9241  max mem: 2905
Test:  [20/50]  eta: 0:00:46  loss: 3.6901 (3.7306)  acc1: 29.6000 (30.4381)  acc5: 52.0000 (53.6381)  time: 1.0347  data: 1.0151  max mem: 2905
Test:  [30/50]  eta: 0:00:27  loss: 3.7869 (3.7748)  acc1: 27.2000 (29.9097)  acc5: 50.4000 (52.5936)  time: 1.0609  data: 1.0409  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 3.9946 (3.8499)  acc1: 27.2000 (29.4439)  acc5: 48.8000 (51.5122)  time: 0.8709  data: 0.8508  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 3.7481 (3.7961)  acc1: 29.6000 (29.9840)  acc5: 51.2000 (52.3680)  time: 0.7295  data: 0.7082  max mem: 2905
Test: Total time: 0:00:51 (1.0237 s / it)
* Acc@1 30.328 Acc@5 52.902 loss 3.751
Accuracy of the model on the 50000 test images: 30.3%
Max accuracy: 50.54%
Epoch: [147]  [  0/625]  eta: 3:58:55  lr: 0.002291  min_lr: 0.002291  loss: 3.3517 (3.3517)  class_acc: 0.4922 (0.4922)  weight_decay: 0.0500 (0.0500)  time: 22.9374  data: 18.6722  max mem: 2905
Epoch: [147]  [200/625]  eta: 0:14:45  lr: 0.002284  min_lr: 0.002284  loss: 3.0038 (3.0383)  class_acc: 0.5156 (0.5182)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1349 (2.4594)  time: 1.9008  data: 0.0013  max mem: 2905
Epoch: [147]  [400/625]  eta: 0:07:32  lr: 0.002277  min_lr: 0.002277  loss: 3.0456 (3.0355)  class_acc: 0.5156 (0.5176)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.4006 (2.5693)  time: 1.9305  data: 0.0008  max mem: 2905
Epoch: [147]  [600/625]  eta: 0:00:50  lr: 0.002270  min_lr: 0.002270  loss: 3.0956 (3.0467)  class_acc: 0.4961 (0.5152)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1714 (2.5219)  time: 1.9798  data: 0.0009  max mem: 2905
Epoch: [147]  [624/625]  eta: 0:00:01  lr: 0.002269  min_lr: 0.002269  loss: 3.0867 (3.0477)  class_acc: 0.5078 (0.5150)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8592 (2.5137)  time: 0.7221  data: 0.0020  max mem: 2905
Epoch: [147] Total time: 0:20:22 (1.9562 s / it)
Averaged stats: lr: 0.002269  min_lr: 0.002269  loss: 3.0867 (3.0508)  class_acc: 0.5078 (0.5149)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8592 (2.5137)
Test:  [ 0/50]  eta: 0:10:51  loss: 2.5123 (2.5123)  acc1: 46.4000 (46.4000)  acc5: 73.6000 (73.6000)  time: 13.0329  data: 13.0048  max mem: 2905
Test:  [10/50]  eta: 0:01:27  loss: 2.5123 (2.5746)  acc1: 47.2000 (46.3273)  acc5: 73.6000 (71.5636)  time: 2.1962  data: 2.1753  max mem: 2905
Test:  [20/50]  eta: 0:00:51  loss: 2.7364 (2.7569)  acc1: 44.8000 (43.5429)  acc5: 68.0000 (68.5333)  time: 1.1338  data: 1.1129  max mem: 2905
Test:  [30/50]  eta: 0:00:29  loss: 2.9157 (2.8039)  acc1: 40.8000 (42.8903)  acc5: 66.4000 (68.1806)  time: 1.0853  data: 1.0654  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 2.9157 (2.8110)  acc1: 40.8000 (42.8293)  acc5: 66.4000 (67.8829)  time: 0.7223  data: 0.7038  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.6534 (2.7788)  acc1: 41.6000 (43.0880)  acc5: 67.2000 (68.3200)  time: 0.6628  data: 0.6435  max mem: 2905
Test: Total time: 0:00:50 (1.0193 s / it)
* Acc@1 43.650 Acc@5 68.506 loss 2.761
Accuracy of the model on the 50000 test images: 43.7%
Max accuracy: 50.54%
Epoch: [148]  [  0/625]  eta: 4:09:23  lr: 0.002269  min_lr: 0.002269  loss: 3.1695 (3.1695)  class_acc: 0.4688 (0.4688)  weight_decay: 0.0500 (0.0500)  time: 23.9415  data: 18.5702  max mem: 2905
Epoch: [148]  [200/625]  eta: 0:14:16  lr: 0.002262  min_lr: 0.002262  loss: 2.9880 (3.0238)  class_acc: 0.5273 (0.5188)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4833 (2.5539)  time: 1.8423  data: 0.5251  max mem: 2905
Epoch: [148]  [400/625]  eta: 0:07:19  lr: 0.002255  min_lr: 0.002255  loss: 3.0093 (3.0311)  class_acc: 0.5352 (0.5189)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4527 (2.5450)  time: 1.9003  data: 0.0109  max mem: 2905
Epoch: [148]  [600/625]  eta: 0:00:48  lr: 0.002248  min_lr: 0.002248  loss: 3.0577 (3.0411)  class_acc: 0.5156 (0.5168)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7720 (2.5787)  time: 1.9974  data: 0.3803  max mem: 2905
Epoch: [148]  [624/625]  eta: 0:00:01  lr: 0.002247  min_lr: 0.002247  loss: 3.0280 (3.0412)  class_acc: 0.5273 (0.5168)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8916 (2.5543)  time: 1.1760  data: 0.0593  max mem: 2905
Epoch: [148] Total time: 0:19:42 (1.8920 s / it)
Averaged stats: lr: 0.002247  min_lr: 0.002247  loss: 3.0280 (3.0448)  class_acc: 0.5273 (0.5164)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8916 (2.5543)
Test:  [ 0/50]  eta: 0:09:52  loss: 2.4585 (2.4585)  acc1: 41.6000 (41.6000)  acc5: 74.4000 (74.4000)  time: 11.8444  data: 11.8086  max mem: 2905
Test:  [10/50]  eta: 0:01:26  loss: 2.8436 (2.8866)  acc1: 41.6000 (41.6727)  acc5: 65.6000 (65.7455)  time: 2.1555  data: 2.1350  max mem: 2905
Test:  [20/50]  eta: 0:00:51  loss: 3.0068 (2.9908)  acc1: 40.8000 (40.2286)  acc5: 62.4000 (64.2667)  time: 1.2254  data: 1.2065  max mem: 2905
Test:  [30/50]  eta: 0:00:28  loss: 3.0115 (2.9515)  acc1: 40.8000 (41.0839)  acc5: 63.2000 (65.1097)  time: 0.9994  data: 0.9798  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 2.9922 (2.9965)  acc1: 41.6000 (40.6439)  acc5: 63.2000 (64.7610)  time: 0.5585  data: 0.5398  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 2.8542 (2.9806)  acc1: 41.6000 (40.6400)  acc5: 65.6000 (65.0080)  time: 0.4895  data: 0.4708  max mem: 2905
Test: Total time: 0:00:48 (0.9743 s / it)
* Acc@1 40.844 Acc@5 65.812 loss 2.945
Accuracy of the model on the 50000 test images: 40.8%
Max accuracy: 50.54%
Epoch: [149]  [  0/625]  eta: 3:55:14  lr: 0.002247  min_lr: 0.002247  loss: 3.1467 (3.1467)  class_acc: 0.4570 (0.4570)  weight_decay: 0.0500 (0.0500)  time: 22.5826  data: 18.4608  max mem: 2905
Epoch: [149]  [200/625]  eta: 0:13:40  lr: 0.002240  min_lr: 0.002240  loss: 3.0379 (3.0399)  class_acc: 0.5195 (0.5166)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.6151 (2.3642)  time: 1.6730  data: 0.8320  max mem: 2905
Epoch: [149]  [400/625]  eta: 0:07:13  lr: 0.002232  min_lr: 0.002232  loss: 2.9688 (3.0396)  class_acc: 0.5312 (0.5178)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3880 (2.6806)  time: 1.9757  data: 0.0060  max mem: 2905
Epoch: [149]  [600/625]  eta: 0:00:49  lr: 0.002225  min_lr: 0.002225  loss: 3.0846 (3.0494)  class_acc: 0.4961 (0.5148)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4737 (2.6573)  time: 2.1015  data: 0.0458  max mem: 2905
Epoch: [149]  [624/625]  eta: 0:00:01  lr: 0.002224  min_lr: 0.002224  loss: 3.0867 (3.0503)  class_acc: 0.5000 (0.5144)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3000 (2.6448)  time: 0.7234  data: 0.0021  max mem: 2905
Epoch: [149] Total time: 0:20:16 (1.9466 s / it)
Averaged stats: lr: 0.002224  min_lr: 0.002224  loss: 3.0867 (3.0456)  class_acc: 0.5000 (0.5158)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3000 (2.6448)
Test:  [ 0/50]  eta: 0:11:19  loss: 2.5075 (2.5075)  acc1: 38.4000 (38.4000)  acc5: 70.4000 (70.4000)  time: 13.5836  data: 13.5596  max mem: 2905
Test:  [10/50]  eta: 0:01:34  loss: 2.6579 (2.7780)  acc1: 40.0000 (42.9091)  acc5: 68.8000 (67.1273)  time: 2.3603  data: 2.3389  max mem: 2905
Test:  [20/50]  eta: 0:00:56  loss: 2.9266 (2.9420)  acc1: 38.4000 (39.9238)  acc5: 64.8000 (65.2191)  time: 1.2932  data: 1.2714  max mem: 2905
Test:  [30/50]  eta: 0:00:33  loss: 3.0047 (2.9357)  acc1: 36.8000 (40.4645)  acc5: 64.8000 (65.4452)  time: 1.2669  data: 1.2429  max mem: 2905
Test:  [40/50]  eta: 0:00:13  loss: 2.8491 (2.9547)  acc1: 43.2000 (40.5659)  acc5: 66.4000 (65.0341)  time: 0.8628  data: 0.8400  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.8092 (2.9274)  acc1: 43.2000 (40.8800)  acc5: 68.8000 (65.6800)  time: 0.8604  data: 0.8399  max mem: 2905
Test: Total time: 0:00:57 (1.1413 s / it)
* Acc@1 41.758 Acc@5 66.238 loss 2.885
Accuracy of the model on the 50000 test images: 41.8%
Max accuracy: 50.54%
Epoch: [150]  [  0/625]  eta: 3:48:53  lr: 0.002224  min_lr: 0.002224  loss: 3.0979 (3.0979)  class_acc: 0.4961 (0.4961)  weight_decay: 0.0500 (0.0500)  time: 21.9732  data: 20.1845  max mem: 2905
Epoch: [150]  [200/625]  eta: 0:14:54  lr: 0.002217  min_lr: 0.002217  loss: 3.0001 (3.0312)  class_acc: 0.5312 (0.5190)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2735 (2.5337)  time: 2.1304  data: 1.9446  max mem: 2905
Epoch: [150]  [400/625]  eta: 0:07:34  lr: 0.002210  min_lr: 0.002210  loss: 3.0345 (3.0356)  class_acc: 0.5039 (0.5184)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4768 (2.5608)  time: 1.8880  data: 1.6795  max mem: 2905
Epoch: [150]  [600/625]  eta: 0:00:49  lr: 0.002203  min_lr: 0.002203  loss: 3.0437 (3.0402)  class_acc: 0.5273 (0.5181)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7446 (2.5425)  time: 1.9584  data: 1.7877  max mem: 2905
Epoch: [150]  [624/625]  eta: 0:00:01  lr: 0.002202  min_lr: 0.002202  loss: 3.0401 (3.0396)  class_acc: 0.5156 (0.5183)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7496 (2.5655)  time: 0.7883  data: 0.6352  max mem: 2905
Epoch: [150] Total time: 0:20:17 (1.9480 s / it)
Averaged stats: lr: 0.002202  min_lr: 0.002202  loss: 3.0401 (3.0419)  class_acc: 0.5156 (0.5165)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7496 (2.5655)
Test:  [ 0/50]  eta: 0:09:40  loss: 2.9424 (2.9424)  acc1: 36.8000 (36.8000)  acc5: 60.0000 (60.0000)  time: 11.6073  data: 11.5793  max mem: 2905
Test:  [10/50]  eta: 0:01:15  loss: 2.7706 (2.8128)  acc1: 40.0000 (41.3091)  acc5: 68.8000 (67.0545)  time: 1.8958  data: 1.8751  max mem: 2905
Test:  [20/50]  eta: 0:00:44  loss: 3.1150 (3.0830)  acc1: 37.6000 (37.9048)  acc5: 64.0000 (63.7333)  time: 0.9640  data: 0.9454  max mem: 2905
Test:  [30/50]  eta: 0:00:25  loss: 3.1970 (3.1249)  acc1: 35.2000 (37.6516)  acc5: 61.6000 (63.1484)  time: 0.9309  data: 0.9132  max mem: 2905
Test:  [40/50]  eta: 0:00:10  loss: 3.1618 (3.1494)  acc1: 36.8000 (37.5610)  acc5: 60.0000 (62.3024)  time: 0.6551  data: 0.6365  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 3.1191 (3.1599)  acc1: 36.8000 (37.3120)  acc5: 60.0000 (62.2400)  time: 0.5938  data: 0.5751  max mem: 2905
Test: Total time: 0:00:46 (0.9210 s / it)
* Acc@1 38.200 Acc@5 62.776 loss 3.120
Accuracy of the model on the 50000 test images: 38.2%
Max accuracy: 50.54%
Epoch: [151]  [  0/625]  eta: 3:44:37  lr: 0.002202  min_lr: 0.002202  loss: 3.1651 (3.1651)  class_acc: 0.5156 (0.5156)  weight_decay: 0.0500 (0.0500)  time: 21.5638  data: 18.6684  max mem: 2905
Epoch: [151]  [200/625]  eta: 0:13:39  lr: 0.002195  min_lr: 0.002195  loss: 3.0265 (3.0355)  class_acc: 0.5039 (0.5178)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2292 (2.4239)  time: 1.7984  data: 0.0006  max mem: 2905
Epoch: [151]  [400/625]  eta: 0:07:08  lr: 0.002188  min_lr: 0.002188  loss: 3.0176 (3.0361)  class_acc: 0.5117 (0.5185)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.0534 (2.6388)  time: 1.7462  data: 0.0007  max mem: 2905
Epoch: [151]  [600/625]  eta: 0:00:47  lr: 0.002181  min_lr: 0.002181  loss: 3.0618 (3.0399)  class_acc: 0.4961 (0.5178)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7182 (2.6694)  time: 1.9069  data: 0.0006  max mem: 2905
Epoch: [151]  [624/625]  eta: 0:00:01  lr: 0.002180  min_lr: 0.002180  loss: 3.0361 (3.0408)  class_acc: 0.5078 (0.5178)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2722 (2.6563)  time: 0.6331  data: 0.0012  max mem: 2905
Epoch: [151] Total time: 0:19:34 (1.8794 s / it)
Averaged stats: lr: 0.002180  min_lr: 0.002180  loss: 3.0361 (3.0411)  class_acc: 0.5078 (0.5170)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2722 (2.6563)
Test:  [ 0/50]  eta: 0:10:12  loss: 3.5451 (3.5451)  acc1: 31.2000 (31.2000)  acc5: 53.6000 (53.6000)  time: 12.2476  data: 12.2141  max mem: 2905
Test:  [10/50]  eta: 0:01:19  loss: 3.2563 (3.2406)  acc1: 37.6000 (36.5091)  acc5: 60.8000 (60.5091)  time: 1.9833  data: 1.9621  max mem: 2905
Test:  [20/50]  eta: 0:00:45  loss: 3.3185 (3.3582)  acc1: 36.0000 (35.1619)  acc5: 58.4000 (59.4286)  time: 0.9639  data: 0.9432  max mem: 2905
Test:  [30/50]  eta: 0:00:25  loss: 3.3992 (3.3403)  acc1: 36.8000 (35.8452)  acc5: 56.8000 (59.6903)  time: 0.9033  data: 0.8821  max mem: 2905
Test:  [40/50]  eta: 0:00:10  loss: 3.2801 (3.3732)  acc1: 34.4000 (35.3561)  acc5: 58.4000 (59.2976)  time: 0.6404  data: 0.6196  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 3.3992 (3.3611)  acc1: 35.2000 (35.7280)  acc5: 60.0000 (59.6800)  time: 0.5934  data: 0.5734  max mem: 2905
Test: Total time: 0:00:47 (0.9515 s / it)
* Acc@1 35.794 Acc@5 60.112 loss 3.324
Accuracy of the model on the 50000 test images: 35.8%
Max accuracy: 50.54%
Epoch: [152]  [  0/625]  eta: 3:22:29  lr: 0.002180  min_lr: 0.002180  loss: 3.0376 (3.0376)  class_acc: 0.4961 (0.4961)  weight_decay: 0.0500 (0.0500)  time: 19.4386  data: 18.2439  max mem: 2905
Epoch: [152]  [200/625]  eta: 0:13:36  lr: 0.002173  min_lr: 0.002173  loss: 3.0593 (3.0118)  class_acc: 0.5156 (0.5253)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7804 (2.2714)  time: 1.7929  data: 0.0321  max mem: 2905
Epoch: [152]  [400/625]  eta: 0:07:03  lr: 0.002165  min_lr: 0.002165  loss: 3.0556 (3.0211)  class_acc: 0.5117 (0.5226)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0346 (2.4357)  time: 1.8247  data: 0.0008  max mem: 2905
Epoch: [152]  [600/625]  eta: 0:00:47  lr: 0.002158  min_lr: 0.002158  loss: 3.0546 (3.0280)  class_acc: 0.5000 (0.5204)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0263 (inf)  time: 2.1767  data: 0.0008  max mem: 2905
Epoch: [152]  [624/625]  eta: 0:00:01  lr: 0.002157  min_lr: 0.002157  loss: 2.9995 (3.0287)  class_acc: 0.5234 (0.5204)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2997 (inf)  time: 0.6602  data: 0.0016  max mem: 2905
Epoch: [152] Total time: 0:19:32 (1.8758 s / it)
Averaged stats: lr: 0.002157  min_lr: 0.002157  loss: 2.9995 (3.0349)  class_acc: 0.5234 (0.5180)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2997 (inf)
Test:  [ 0/50]  eta: 0:10:09  loss: 3.5093 (3.5093)  acc1: 28.8000 (28.8000)  acc5: 54.4000 (54.4000)  time: 12.1903  data: 12.1593  max mem: 2905
Test:  [10/50]  eta: 0:01:29  loss: 3.1881 (3.2337)  acc1: 40.0000 (38.3273)  acc5: 61.6000 (61.6727)  time: 2.2366  data: 2.2156  max mem: 2905
Test:  [20/50]  eta: 0:00:54  loss: 3.3766 (3.3887)  acc1: 35.2000 (35.4286)  acc5: 60.8000 (59.1619)  time: 1.3064  data: 1.2866  max mem: 2905
Test:  [30/50]  eta: 0:00:32  loss: 3.4955 (3.3873)  acc1: 35.2000 (36.0774)  acc5: 56.0000 (58.7613)  time: 1.2908  data: 1.2724  max mem: 2905
Test:  [40/50]  eta: 0:00:13  loss: 3.4960 (3.4144)  acc1: 36.0000 (35.8244)  acc5: 56.0000 (58.3805)  time: 0.8609  data: 0.8432  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 3.5112 (3.4276)  acc1: 33.6000 (35.5840)  acc5: 56.8000 (58.3200)  time: 0.7874  data: 0.7694  max mem: 2905
Test: Total time: 0:00:56 (1.1271 s / it)
* Acc@1 35.734 Acc@5 58.714 loss 3.402
Accuracy of the model on the 50000 test images: 35.7%
Max accuracy: 50.54%
Epoch: [153]  [  0/625]  eta: 3:46:03  lr: 0.002157  min_lr: 0.002157  loss: 2.9352 (2.9352)  class_acc: 0.5039 (0.5039)  weight_decay: 0.0500 (0.0500)  time: 21.7019  data: 21.5852  max mem: 2905
Epoch: [153]  [200/625]  eta: 0:14:39  lr: 0.002150  min_lr: 0.002150  loss: 3.0729 (3.0122)  class_acc: 0.5039 (0.5230)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2612 (2.4603)  time: 1.9413  data: 0.0007  max mem: 2905
Epoch: [153]  [400/625]  eta: 0:07:28  lr: 0.002143  min_lr: 0.002143  loss: 2.9897 (3.0248)  class_acc: 0.5234 (0.5201)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9767 (2.5491)  time: 1.9214  data: 0.0009  max mem: 2905
Epoch: [153]  [600/625]  eta: 0:00:49  lr: 0.002136  min_lr: 0.002136  loss: 3.0572 (3.0379)  class_acc: 0.5117 (0.5177)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6070 (2.5075)  time: 1.9138  data: 0.0007  max mem: 2905
Epoch: [153]  [624/625]  eta: 0:00:01  lr: 0.002135  min_lr: 0.002135  loss: 3.0697 (3.0391)  class_acc: 0.4922 (0.5171)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6070 (2.5045)  time: 0.8963  data: 0.0018  max mem: 2905
Epoch: [153] Total time: 0:20:05 (1.9293 s / it)
Averaged stats: lr: 0.002135  min_lr: 0.002135  loss: 3.0697 (3.0323)  class_acc: 0.4922 (0.5188)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6070 (2.5045)
Test:  [ 0/50]  eta: 0:09:12  loss: 3.2503 (3.2503)  acc1: 37.6000 (37.6000)  acc5: 57.6000 (57.6000)  time: 11.0546  data: 11.0247  max mem: 2905
Test:  [10/50]  eta: 0:01:18  loss: 3.2267 (3.2357)  acc1: 35.2000 (36.5818)  acc5: 64.8000 (62.1818)  time: 1.9732  data: 1.9530  max mem: 2905
Test:  [20/50]  eta: 0:00:47  loss: 3.3884 (3.4245)  acc1: 34.4000 (33.8286)  acc5: 57.6000 (58.7810)  time: 1.1164  data: 1.0975  max mem: 2905
Test:  [30/50]  eta: 0:00:27  loss: 3.5025 (3.3792)  acc1: 32.8000 (35.4839)  acc5: 56.8000 (59.2000)  time: 1.0619  data: 1.0435  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 3.3892 (3.3861)  acc1: 36.0000 (35.7268)  acc5: 57.6000 (59.4732)  time: 0.6433  data: 0.6235  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 3.3892 (3.3864)  acc1: 34.4000 (35.3600)  acc5: 59.2000 (59.5200)  time: 0.5689  data: 0.5490  max mem: 2905
Test: Total time: 0:00:46 (0.9319 s / it)
* Acc@1 36.286 Acc@5 60.354 loss 3.330
Accuracy of the model on the 50000 test images: 36.3%
Max accuracy: 50.54%
Epoch: [154]  [  0/625]  eta: 3:27:30  lr: 0.002135  min_lr: 0.002135  loss: 2.9735 (2.9735)  class_acc: 0.5430 (0.5430)  weight_decay: 0.0500 (0.0500)  time: 19.9212  data: 18.6431  max mem: 2905
Epoch: [154]  [200/625]  eta: 0:13:48  lr: 0.002128  min_lr: 0.002128  loss: 2.9991 (3.0093)  class_acc: 0.5195 (0.5235)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.8436 (2.6363)  time: 1.9476  data: 0.9274  max mem: 2905
Epoch: [154]  [400/625]  eta: 0:07:12  lr: 0.002121  min_lr: 0.002121  loss: 3.0310 (3.0143)  class_acc: 0.5234 (0.5220)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5752 (2.5978)  time: 1.8642  data: 0.3912  max mem: 2905
Epoch: [154]  [600/625]  eta: 0:00:48  lr: 0.002113  min_lr: 0.002113  loss: 3.0270 (3.0280)  class_acc: 0.5117 (0.5194)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.0360 (2.5763)  time: 1.9887  data: 0.0515  max mem: 2905
Epoch: [154]  [624/625]  eta: 0:00:01  lr: 0.002113  min_lr: 0.002113  loss: 3.0285 (3.0280)  class_acc: 0.5156 (0.5194)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.8794 (2.5640)  time: 0.9190  data: 0.0265  max mem: 2905
Epoch: [154] Total time: 0:19:35 (1.8808 s / it)
Averaged stats: lr: 0.002113  min_lr: 0.002113  loss: 3.0285 (3.0267)  class_acc: 0.5156 (0.5197)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.8794 (2.5640)
Test:  [ 0/50]  eta: 0:09:31  loss: 3.8801 (3.8801)  acc1: 28.8000 (28.8000)  acc5: 47.2000 (47.2000)  time: 11.4247  data: 11.3942  max mem: 2905
Test:  [10/50]  eta: 0:01:25  loss: 3.2167 (3.3366)  acc1: 35.2000 (34.9818)  acc5: 60.0000 (58.6909)  time: 2.1347  data: 2.1136  max mem: 2905
Test:  [20/50]  eta: 0:00:51  loss: 3.4895 (3.4943)  acc1: 32.8000 (32.6476)  acc5: 56.0000 (56.7238)  time: 1.2394  data: 1.2195  max mem: 2905
Test:  [30/50]  eta: 0:00:29  loss: 3.6375 (3.4702)  acc1: 29.6000 (33.0581)  acc5: 55.2000 (56.9806)  time: 1.0920  data: 1.0730  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 3.5073 (3.5147)  acc1: 32.8000 (32.3902)  acc5: 56.0000 (56.5463)  time: 0.6334  data: 0.6154  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 3.4358 (3.4976)  acc1: 33.6000 (32.8960)  acc5: 57.6000 (56.8640)  time: 0.5545  data: 0.5366  max mem: 2905
Test: Total time: 0:00:49 (0.9901 s / it)
* Acc@1 33.546 Acc@5 57.298 loss 3.466
Accuracy of the model on the 50000 test images: 33.5%
Max accuracy: 50.54%
Epoch: [155]  [  0/625]  eta: 3:47:09  lr: 0.002113  min_lr: 0.002113  loss: 2.7752 (2.7752)  class_acc: 0.5625 (0.5625)  weight_decay: 0.0500 (0.0500)  time: 21.8076  data: 18.6507  max mem: 2905
Epoch: [155]  [200/625]  eta: 0:13:55  lr: 0.002105  min_lr: 0.002105  loss: 3.0109 (3.0089)  class_acc: 0.5156 (0.5243)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7185 (2.7366)  time: 1.8816  data: 0.0012  max mem: 2905
Epoch: [155]  [400/625]  eta: 0:07:20  lr: 0.002098  min_lr: 0.002098  loss: 2.9925 (3.0268)  class_acc: 0.5156 (0.5199)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5118 (2.6630)  time: 1.9407  data: 0.0010  max mem: 2905
Epoch: [155]  [600/625]  eta: 0:00:49  lr: 0.002091  min_lr: 0.002091  loss: 3.0713 (3.0340)  class_acc: 0.5117 (0.5181)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7920 (2.6555)  time: 1.9477  data: 0.0483  max mem: 2905
Epoch: [155]  [624/625]  eta: 0:00:01  lr: 0.002090  min_lr: 0.002090  loss: 3.0200 (3.0344)  class_acc: 0.5117 (0.5181)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0642 (2.6587)  time: 0.5837  data: 0.0060  max mem: 2905
Epoch: [155] Total time: 0:20:13 (1.9418 s / it)
Averaged stats: lr: 0.002090  min_lr: 0.002090  loss: 3.0200 (3.0307)  class_acc: 0.5117 (0.5197)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0642 (2.6587)
Test:  [ 0/50]  eta: 0:10:44  loss: 3.4774 (3.4774)  acc1: 34.4000 (34.4000)  acc5: 55.2000 (55.2000)  time: 12.8984  data: 12.8685  max mem: 2905
Test:  [10/50]  eta: 0:01:25  loss: 2.8554 (2.9839)  acc1: 40.8000 (40.2182)  acc5: 68.0000 (65.0909)  time: 2.1309  data: 2.1080  max mem: 2905
Test:  [20/50]  eta: 0:00:49  loss: 3.2216 (3.2589)  acc1: 36.0000 (36.1143)  acc5: 60.0000 (60.7619)  time: 1.0896  data: 1.0685  max mem: 2905
Test:  [30/50]  eta: 0:00:29  loss: 3.3724 (3.2594)  acc1: 34.4000 (35.7419)  acc5: 56.8000 (60.3613)  time: 1.1041  data: 1.0842  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 3.3724 (3.3082)  acc1: 34.4000 (35.0244)  acc5: 56.8000 (59.9610)  time: 0.9255  data: 0.9064  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 3.3747 (3.3033)  acc1: 32.0000 (35.2160)  acc5: 56.8000 (60.0640)  time: 0.8749  data: 0.8562  max mem: 2905
Test: Total time: 0:00:55 (1.1094 s / it)
* Acc@1 35.538 Acc@5 60.466 loss 3.269
Accuracy of the model on the 50000 test images: 35.5%
Max accuracy: 50.54%
Epoch: [156]  [  0/625]  eta: 3:47:19  lr: 0.002090  min_lr: 0.002090  loss: 2.9601 (2.9601)  class_acc: 0.5273 (0.5273)  weight_decay: 0.0500 (0.0500)  time: 21.8231  data: 18.2878  max mem: 2905
Epoch: [156]  [200/625]  eta: 0:14:40  lr: 0.002083  min_lr: 0.002083  loss: 2.9491 (3.0054)  class_acc: 0.5195 (0.5225)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4504 (2.4339)  time: 1.8751  data: 0.0008  max mem: 2905
Epoch: [156]  [400/625]  eta: 0:07:33  lr: 0.002076  min_lr: 0.002076  loss: 3.0714 (3.0159)  class_acc: 0.5039 (0.5201)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3515 (2.4555)  time: 1.9844  data: 0.0007  max mem: 2905
Epoch: [156]  [600/625]  eta: 0:00:49  lr: 0.002069  min_lr: 0.002069  loss: 3.0238 (3.0244)  class_acc: 0.5195 (0.5186)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.6113 (2.4710)  time: 2.0847  data: 0.0007  max mem: 2905
Epoch: [156]  [624/625]  eta: 0:00:01  lr: 0.002068  min_lr: 0.002068  loss: 3.0534 (3.0250)  class_acc: 0.5195 (0.5186)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3017 (2.4959)  time: 0.7239  data: 0.0013  max mem: 2905
Epoch: [156] Total time: 0:20:12 (1.9398 s / it)
Averaged stats: lr: 0.002068  min_lr: 0.002068  loss: 3.0534 (3.0243)  class_acc: 0.5195 (0.5201)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3017 (2.4959)
Test:  [ 0/50]  eta: 0:10:22  loss: 2.9412 (2.9412)  acc1: 40.0000 (40.0000)  acc5: 66.4000 (66.4000)  time: 12.4532  data: 12.4195  max mem: 2905
Test:  [10/50]  eta: 0:01:23  loss: 3.0171 (3.1125)  acc1: 40.0000 (39.7091)  acc5: 62.4000 (61.3818)  time: 2.0921  data: 2.0707  max mem: 2905
Test:  [20/50]  eta: 0:00:48  loss: 3.2026 (3.3091)  acc1: 36.0000 (36.4952)  acc5: 59.2000 (59.1238)  time: 1.0749  data: 1.0554  max mem: 2905
Test:  [30/50]  eta: 0:00:28  loss: 3.5017 (3.3518)  acc1: 33.6000 (35.9742)  acc5: 58.4000 (58.7355)  time: 1.0429  data: 1.0242  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 3.3821 (3.3553)  acc1: 35.2000 (35.6683)  acc5: 58.4000 (58.7902)  time: 0.6676  data: 0.6497  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 3.3477 (3.3675)  acc1: 34.4000 (35.2480)  acc5: 59.2000 (58.6880)  time: 0.6143  data: 0.5965  max mem: 2905
Test: Total time: 0:00:48 (0.9630 s / it)
* Acc@1 35.434 Acc@5 58.678 loss 3.355
Accuracy of the model on the 50000 test images: 35.4%
Max accuracy: 50.54%
Epoch: [157]  [  0/625]  eta: 3:36:36  lr: 0.002068  min_lr: 0.002068  loss: 3.1194 (3.1194)  class_acc: 0.4922 (0.4922)  weight_decay: 0.0500 (0.0500)  time: 20.7952  data: 19.5208  max mem: 2905
Epoch: [157]  [200/625]  eta: 0:14:09  lr: 0.002061  min_lr: 0.002061  loss: 3.0206 (3.0142)  class_acc: 0.5273 (0.5254)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5901 (3.0417)  time: 1.8902  data: 0.0017  max mem: 2905
Epoch: [157]  [400/625]  eta: 0:07:21  lr: 0.002053  min_lr: 0.002053  loss: 3.0223 (3.0161)  class_acc: 0.5195 (0.5234)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1142 (2.6565)  time: 1.9916  data: 0.0007  max mem: 2905
Epoch: [157]  [600/625]  eta: 0:00:49  lr: 0.002046  min_lr: 0.002046  loss: 3.0597 (3.0238)  class_acc: 0.5078 (0.5220)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8276 (2.7122)  time: 2.1000  data: 0.0006  max mem: 2905
Epoch: [157]  [624/625]  eta: 0:00:01  lr: 0.002045  min_lr: 0.002045  loss: 3.0107 (3.0244)  class_acc: 0.5156 (0.5217)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3966 (2.7331)  time: 0.8472  data: 0.0020  max mem: 2905
Epoch: [157] Total time: 0:20:12 (1.9402 s / it)
Averaged stats: lr: 0.002045  min_lr: 0.002045  loss: 3.0107 (3.0219)  class_acc: 0.5156 (0.5211)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3966 (2.7331)
Test:  [ 0/50]  eta: 0:10:56  loss: 2.8063 (2.8063)  acc1: 40.8000 (40.8000)  acc5: 65.6000 (65.6000)  time: 13.1270  data: 13.0957  max mem: 2905
Test:  [10/50]  eta: 0:01:28  loss: 2.8591 (2.9528)  acc1: 40.8000 (41.1636)  acc5: 68.0000 (66.2545)  time: 2.2014  data: 2.1818  max mem: 2905
Test:  [20/50]  eta: 0:00:53  loss: 3.0065 (3.0680)  acc1: 38.4000 (38.7429)  acc5: 64.8000 (63.8095)  time: 1.2065  data: 1.1871  max mem: 2905
Test:  [30/50]  eta: 0:00:31  loss: 3.0294 (3.0176)  acc1: 37.6000 (39.9484)  acc5: 62.4000 (64.6194)  time: 1.2598  data: 1.2409  max mem: 2905
Test:  [40/50]  eta: 0:00:13  loss: 3.0294 (3.0609)  acc1: 36.8000 (39.0049)  acc5: 62.4000 (63.9220)  time: 0.9770  data: 0.9591  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 3.0332 (3.0391)  acc1: 36.8000 (38.9600)  acc5: 63.2000 (64.4640)  time: 0.9116  data: 0.8909  max mem: 2905
Test: Total time: 0:00:57 (1.1489 s / it)
* Acc@1 39.884 Acc@5 65.088 loss 2.998
Accuracy of the model on the 50000 test images: 39.9%
Max accuracy: 50.54%
Epoch: [158]  [  0/625]  eta: 3:50:33  lr: 0.002045  min_lr: 0.002045  loss: 2.9035 (2.9035)  class_acc: 0.5195 (0.5195)  weight_decay: 0.0500 (0.0500)  time: 22.1338  data: 22.0038  max mem: 2905
Epoch: [158]  [200/625]  eta: 0:14:31  lr: 0.002038  min_lr: 0.002038  loss: 3.0524 (3.0105)  class_acc: 0.5078 (0.5228)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7281 (2.8381)  time: 1.9043  data: 0.9828  max mem: 2905
Epoch: [158]  [400/625]  eta: 0:07:25  lr: 0.002031  min_lr: 0.002031  loss: 3.0553 (3.0218)  class_acc: 0.5117 (0.5199)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0558 (2.7807)  time: 2.0683  data: 0.0006  max mem: 2905
Epoch: [158]  [600/625]  eta: 0:00:49  lr: 0.002024  min_lr: 0.002024  loss: 3.0205 (3.0309)  class_acc: 0.5273 (0.5187)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8804 (2.6360)  time: 2.1269  data: 0.0008  max mem: 2905
Epoch: [158]  [624/625]  eta: 0:00:01  lr: 0.002023  min_lr: 0.002023  loss: 3.0348 (3.0309)  class_acc: 0.5117 (0.5184)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6145 (2.6381)  time: 0.8863  data: 0.0101  max mem: 2905
Epoch: [158] Total time: 0:20:18 (1.9494 s / it)
Averaged stats: lr: 0.002023  min_lr: 0.002023  loss: 3.0348 (3.0214)  class_acc: 0.5117 (0.5216)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6145 (2.6381)
Test:  [ 0/50]  eta: 0:10:51  loss: 2.8790 (2.8790)  acc1: 42.4000 (42.4000)  acc5: 64.8000 (64.8000)  time: 13.0309  data: 13.0038  max mem: 2905
Test:  [10/50]  eta: 0:01:29  loss: 2.8306 (2.7522)  acc1: 44.8000 (43.3455)  acc5: 64.8000 (66.9091)  time: 2.2299  data: 2.2101  max mem: 2905
Test:  [20/50]  eta: 0:00:52  loss: 3.0101 (2.9438)  acc1: 38.4000 (40.1524)  acc5: 64.0000 (65.3714)  time: 1.1820  data: 1.1611  max mem: 2905
Test:  [30/50]  eta: 0:00:30  loss: 3.1355 (2.9555)  acc1: 37.6000 (40.7226)  acc5: 64.8000 (65.1613)  time: 1.1386  data: 1.1167  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 3.0745 (2.9753)  acc1: 39.2000 (40.3707)  acc5: 65.6000 (64.7024)  time: 0.7614  data: 0.7422  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 3.0783 (3.0112)  acc1: 37.6000 (39.9200)  acc5: 61.6000 (64.3520)  time: 0.7104  data: 0.6923  max mem: 2905
Test: Total time: 0:00:52 (1.0555 s / it)
* Acc@1 40.394 Acc@5 64.784 loss 2.969
Accuracy of the model on the 50000 test images: 40.4%
Max accuracy: 50.54%
Epoch: [159]  [  0/625]  eta: 3:47:21  lr: 0.002023  min_lr: 0.002023  loss: 3.0250 (3.0250)  class_acc: 0.4844 (0.4844)  weight_decay: 0.0500 (0.0500)  time: 21.8270  data: 18.4449  max mem: 2905
Epoch: [159]  [200/625]  eta: 0:14:21  lr: 0.002016  min_lr: 0.002016  loss: 3.0319 (2.9930)  class_acc: 0.5234 (0.5266)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3921 (2.4452)  time: 1.8829  data: 0.3644  max mem: 2905
Epoch: [159]  [400/625]  eta: 0:07:29  lr: 0.002009  min_lr: 0.002009  loss: 3.0498 (3.0097)  class_acc: 0.5156 (0.5233)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4559 (inf)  time: 2.0620  data: 0.0007  max mem: 2905
Epoch: [159]  [600/625]  eta: 0:00:49  lr: 0.002001  min_lr: 0.002001  loss: 3.0302 (3.0178)  class_acc: 0.5117 (0.5224)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.4642 (inf)  time: 1.9790  data: 0.0007  max mem: 2905
Epoch: [159]  [624/625]  eta: 0:00:01  lr: 0.002001  min_lr: 0.002001  loss: 3.0329 (3.0196)  class_acc: 0.5000 (0.5218)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6362 (inf)  time: 1.1082  data: 0.0014  max mem: 2905
Epoch: [159] Total time: 0:20:16 (1.9458 s / it)
Averaged stats: lr: 0.002001  min_lr: 0.002001  loss: 3.0329 (3.0174)  class_acc: 0.5000 (0.5229)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6362 (inf)
Test:  [ 0/50]  eta: 0:10:02  loss: 2.7988 (2.7988)  acc1: 43.2000 (43.2000)  acc5: 69.6000 (69.6000)  time: 12.0570  data: 12.0300  max mem: 2905
Test:  [10/50]  eta: 0:01:22  loss: 3.0871 (3.0594)  acc1: 36.8000 (38.9091)  acc5: 64.0000 (63.6364)  time: 2.0502  data: 2.0272  max mem: 2905
Test:  [20/50]  eta: 0:00:49  loss: 3.1623 (3.1333)  acc1: 36.8000 (37.0667)  acc5: 62.4000 (62.7810)  time: 1.1139  data: 1.0922  max mem: 2905
Test:  [30/50]  eta: 0:00:27  loss: 3.1623 (3.1303)  acc1: 36.0000 (37.8323)  acc5: 62.4000 (62.6065)  time: 1.0101  data: 0.9899  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 3.0786 (3.1501)  acc1: 38.4000 (37.6781)  acc5: 61.6000 (62.1268)  time: 0.5755  data: 0.5549  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 3.1692 (3.1886)  acc1: 36.0000 (37.1840)  acc5: 61.6000 (61.7120)  time: 0.4883  data: 0.4684  max mem: 2905
Test: Total time: 0:00:46 (0.9363 s / it)
* Acc@1 38.046 Acc@5 62.350 loss 3.146
Accuracy of the model on the 50000 test images: 38.0%
Max accuracy: 50.54%
Epoch: [160]  [  0/625]  eta: 3:26:54  lr: 0.002001  min_lr: 0.002001  loss: 2.9900 (2.9900)  class_acc: 0.4922 (0.4922)  weight_decay: 0.0500 (0.0500)  time: 19.8626  data: 19.5622  max mem: 2905
Epoch: [160]  [200/625]  eta: 0:13:38  lr: 0.001993  min_lr: 0.001993  loss: 2.9546 (2.9929)  class_acc: 0.5273 (0.5270)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3242 (2.3992)  time: 1.9188  data: 0.0005  max mem: 2905
Epoch: [160]  [400/625]  eta: 0:07:17  lr: 0.001986  min_lr: 0.001986  loss: 3.0093 (3.0115)  class_acc: 0.5156 (0.5231)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7988 (2.5474)  time: 1.9939  data: 0.0007  max mem: 2905
Epoch: [160]  [600/625]  eta: 0:00:49  lr: 0.001979  min_lr: 0.001979  loss: 3.0520 (3.0162)  class_acc: 0.5312 (0.5226)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4405 (2.5678)  time: 2.0670  data: 0.0278  max mem: 2905
Epoch: [160]  [624/625]  eta: 0:00:01  lr: 0.001978  min_lr: 0.001978  loss: 3.0520 (3.0175)  class_acc: 0.5156 (0.5221)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0516 (2.5486)  time: 0.7611  data: 0.0015  max mem: 2905
Epoch: [160] Total time: 0:19:59 (1.9190 s / it)
Averaged stats: lr: 0.001978  min_lr: 0.001978  loss: 3.0520 (3.0129)  class_acc: 0.5156 (0.5229)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0516 (2.5486)
Test:  [ 0/50]  eta: 0:11:26  loss: 2.9394 (2.9394)  acc1: 29.6000 (29.6000)  acc5: 68.8000 (68.8000)  time: 13.7376  data: 13.7101  max mem: 2905
Test:  [10/50]  eta: 0:01:29  loss: 3.0636 (3.0320)  acc1: 40.8000 (38.4000)  acc5: 61.6000 (63.6364)  time: 2.2285  data: 2.2094  max mem: 2905
Test:  [20/50]  eta: 0:00:51  loss: 3.1471 (3.1398)  acc1: 37.6000 (36.6095)  acc5: 61.6000 (62.2095)  time: 1.1312  data: 1.1122  max mem: 2905
Test:  [30/50]  eta: 0:00:30  loss: 3.1471 (3.1000)  acc1: 35.2000 (37.5742)  acc5: 60.0000 (62.2194)  time: 1.1692  data: 1.1500  max mem: 2905
Test:  [40/50]  eta: 0:00:13  loss: 3.0694 (3.1393)  acc1: 37.6000 (37.1122)  acc5: 60.8000 (62.1463)  time: 0.9834  data: 0.9645  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 3.0694 (3.1121)  acc1: 37.6000 (37.3440)  acc5: 64.8000 (62.7520)  time: 0.9017  data: 0.8804  max mem: 2905
Test: Total time: 0:00:56 (1.1384 s / it)
* Acc@1 38.436 Acc@5 63.624 loss 3.063
Accuracy of the model on the 50000 test images: 38.4%
Max accuracy: 50.54%
Epoch: [161]  [  0/625]  eta: 3:32:26  lr: 0.001978  min_lr: 0.001978  loss: 2.9341 (2.9341)  class_acc: 0.5586 (0.5586)  weight_decay: 0.0500 (0.0500)  time: 20.3938  data: 19.5079  max mem: 2905
Epoch: [161]  [200/625]  eta: 0:14:37  lr: 0.001971  min_lr: 0.001971  loss: 3.0436 (3.0027)  class_acc: 0.5195 (0.5256)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0281 (2.5309)  time: 1.8008  data: 0.0010  max mem: 2905
Epoch: [161]  [400/625]  eta: 0:07:30  lr: 0.001964  min_lr: 0.001964  loss: 3.0016 (3.0113)  class_acc: 0.5195 (0.5233)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.9704 (2.5799)  time: 1.9437  data: 0.0008  max mem: 2905
Epoch: [161]  [600/625]  eta: 0:00:49  lr: 0.001956  min_lr: 0.001956  loss: 2.9813 (3.0121)  class_acc: 0.5234 (0.5227)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4052 (2.5571)  time: 2.0764  data: 0.0008  max mem: 2905
Epoch: [161]  [624/625]  eta: 0:00:01  lr: 0.001956  min_lr: 0.001956  loss: 3.0267 (3.0123)  class_acc: 0.5156 (0.5226)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5296 (2.5531)  time: 0.7451  data: 0.0042  max mem: 2905
Epoch: [161] Total time: 0:20:13 (1.9420 s / it)
Averaged stats: lr: 0.001956  min_lr: 0.001956  loss: 3.0267 (3.0114)  class_acc: 0.5156 (0.5236)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5296 (2.5531)
Test:  [ 0/50]  eta: 0:10:19  loss: 3.7432 (3.7432)  acc1: 31.2000 (31.2000)  acc5: 55.2000 (55.2000)  time: 12.3817  data: 12.3492  max mem: 2905
Test:  [10/50]  eta: 0:01:22  loss: 3.2487 (3.2970)  acc1: 36.0000 (35.0545)  acc5: 60.0000 (60.8000)  time: 2.0615  data: 2.0393  max mem: 2905
Test:  [20/50]  eta: 0:00:48  loss: 3.2875 (3.4076)  acc1: 34.4000 (33.6000)  acc5: 59.2000 (58.9714)  time: 1.0720  data: 1.0512  max mem: 2905
Test:  [30/50]  eta: 0:00:27  loss: 3.4490 (3.3851)  acc1: 32.8000 (34.2194)  acc5: 58.4000 (59.4839)  time: 0.9984  data: 0.9786  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 3.3444 (3.4059)  acc1: 32.0000 (34.3415)  acc5: 57.6000 (58.9463)  time: 0.6106  data: 0.5908  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 3.2676 (3.3755)  acc1: 35.2000 (34.6880)  acc5: 57.6000 (59.2800)  time: 0.5106  data: 0.4912  max mem: 2905
Test: Total time: 0:00:47 (0.9414 s / it)
* Acc@1 34.820 Acc@5 59.346 loss 3.352
Accuracy of the model on the 50000 test images: 34.8%
Max accuracy: 50.54%
Epoch: [162]  [  0/625]  eta: 3:39:08  lr: 0.001956  min_lr: 0.001956  loss: 2.9780 (2.9780)  class_acc: 0.5156 (0.5156)  weight_decay: 0.0500 (0.0500)  time: 21.0377  data: 17.2221  max mem: 2905
Epoch: [162]  [200/625]  eta: 0:14:13  lr: 0.001948  min_lr: 0.001948  loss: 3.0291 (2.9952)  class_acc: 0.5156 (0.5262)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.6709 (2.4841)  time: 1.8659  data: 0.0280  max mem: 2905
Epoch: [162]  [400/625]  eta: 0:07:18  lr: 0.001941  min_lr: 0.001941  loss: 2.9941 (2.9947)  class_acc: 0.5273 (0.5274)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5552 (2.5308)  time: 1.9128  data: 0.0053  max mem: 2905
Epoch: [162]  [600/625]  eta: 0:00:48  lr: 0.001934  min_lr: 0.001934  loss: 3.0022 (3.0015)  class_acc: 0.5195 (0.5256)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0132 (2.5374)  time: 1.9884  data: 0.0007  max mem: 2905
Epoch: [162]  [624/625]  eta: 0:00:01  lr: 0.001933  min_lr: 0.001933  loss: 2.9778 (3.0021)  class_acc: 0.5234 (0.5255)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0642 (2.5320)  time: 0.9402  data: 0.0023  max mem: 2905
Epoch: [162] Total time: 0:19:38 (1.8859 s / it)
Averaged stats: lr: 0.001933  min_lr: 0.001933  loss: 2.9778 (3.0061)  class_acc: 0.5234 (0.5243)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0642 (2.5320)
Test:  [ 0/50]  eta: 0:10:10  loss: 2.2307 (2.2307)  acc1: 56.8000 (56.8000)  acc5: 75.2000 (75.2000)  time: 12.2016  data: 12.1716  max mem: 2905
Test:  [10/50]  eta: 0:01:17  loss: 2.3625 (2.3458)  acc1: 54.4000 (51.4182)  acc5: 72.8000 (73.6727)  time: 1.9343  data: 1.9151  max mem: 2905
Test:  [20/50]  eta: 0:00:44  loss: 2.4769 (2.4956)  acc1: 48.0000 (48.3429)  acc5: 72.0000 (72.4571)  time: 0.9447  data: 0.9253  max mem: 2905
Test:  [30/50]  eta: 0:00:25  loss: 2.5498 (2.5496)  acc1: 44.0000 (47.3032)  acc5: 71.2000 (71.6645)  time: 0.9321  data: 0.9123  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 2.6894 (2.5854)  acc1: 45.6000 (47.2195)  acc5: 68.0000 (71.0829)  time: 0.7578  data: 0.7393  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.7985 (2.6443)  acc1: 44.0000 (46.0640)  acc5: 68.0000 (70.5280)  time: 0.7372  data: 0.7183  max mem: 2905
Test: Total time: 0:00:50 (1.0073 s / it)
* Acc@1 46.360 Acc@5 70.986 loss 2.623
Accuracy of the model on the 50000 test images: 46.4%
Max accuracy: 50.54%
Epoch: [163]  [  0/625]  eta: 3:34:21  lr: 0.001933  min_lr: 0.001933  loss: 3.0274 (3.0274)  class_acc: 0.5273 (0.5273)  weight_decay: 0.0500 (0.0500)  time: 20.5791  data: 19.3684  max mem: 2905
Epoch: [163]  [200/625]  eta: 0:13:28  lr: 0.001926  min_lr: 0.001926  loss: 2.9864 (3.0064)  class_acc: 0.5234 (0.5262)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3154 (2.9247)  time: 1.7898  data: 0.0006  max mem: 2905
Epoch: [163]  [400/625]  eta: 0:07:05  lr: 0.001919  min_lr: 0.001919  loss: 2.9711 (2.9996)  class_acc: 0.5312 (0.5268)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1815 (2.7231)  time: 1.7375  data: 0.0006  max mem: 2905
Epoch: [163]  [600/625]  eta: 0:00:47  lr: 0.001912  min_lr: 0.001912  loss: 3.0241 (3.0069)  class_acc: 0.5234 (0.5252)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1012 (2.8291)  time: 1.9009  data: 0.0026  max mem: 2905
Epoch: [163]  [624/625]  eta: 0:00:01  lr: 0.001911  min_lr: 0.001911  loss: 2.9689 (3.0057)  class_acc: 0.5273 (0.5254)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5482 (2.8276)  time: 0.8900  data: 0.0069  max mem: 2905
Epoch: [163] Total time: 0:19:20 (1.8569 s / it)
Averaged stats: lr: 0.001911  min_lr: 0.001911  loss: 2.9689 (3.0062)  class_acc: 0.5273 (0.5245)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5482 (2.8276)
Test:  [ 0/50]  eta: 0:11:30  loss: 2.7785 (2.7785)  acc1: 45.6000 (45.6000)  acc5: 66.4000 (66.4000)  time: 13.8183  data: 13.7918  max mem: 2905
Test:  [10/50]  eta: 0:01:32  loss: 2.7681 (2.8103)  acc1: 44.8000 (45.2364)  acc5: 66.4000 (66.3273)  time: 2.3049  data: 2.2812  max mem: 2905
Test:  [20/50]  eta: 0:00:53  loss: 3.0242 (3.0300)  acc1: 39.2000 (40.8000)  acc5: 64.8000 (64.6857)  time: 1.1824  data: 1.1603  max mem: 2905
Test:  [30/50]  eta: 0:00:29  loss: 3.1845 (3.0246)  acc1: 38.4000 (40.6710)  acc5: 64.0000 (64.4903)  time: 1.0269  data: 1.0061  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 3.0399 (3.0763)  acc1: 39.2000 (40.1951)  acc5: 63.2000 (63.7463)  time: 0.5740  data: 0.5538  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 3.1764 (3.1003)  acc1: 38.4000 (39.8080)  acc5: 62.4000 (63.2960)  time: 0.6283  data: 0.6077  max mem: 2905
Test: Total time: 0:00:50 (1.0091 s / it)
* Acc@1 39.894 Acc@5 63.868 loss 3.059
Accuracy of the model on the 50000 test images: 39.9%
Max accuracy: 50.54%
Epoch: [164]  [  0/625]  eta: 3:48:40  lr: 0.001911  min_lr: 0.001911  loss: 3.0496 (3.0496)  class_acc: 0.5117 (0.5117)  weight_decay: 0.0500 (0.0500)  time: 21.9520  data: 17.7404  max mem: 2905
Epoch: [164]  [200/625]  eta: 0:13:41  lr: 0.001904  min_lr: 0.001904  loss: 2.9989 (2.9934)  class_acc: 0.5273 (0.5261)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5929 (2.7544)  time: 1.8066  data: 0.0006  max mem: 2905
Epoch: [164]  [400/625]  eta: 0:07:08  lr: 0.001896  min_lr: 0.001896  loss: 2.9909 (3.0018)  class_acc: 0.5312 (0.5253)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6663 (2.7015)  time: 1.7716  data: 0.0005  max mem: 2905
Epoch: [164]  [600/625]  eta: 0:00:47  lr: 0.001889  min_lr: 0.001889  loss: 3.0058 (3.0014)  class_acc: 0.5195 (0.5256)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1491 (2.6635)  time: 1.9180  data: 0.0005  max mem: 2905
Epoch: [164]  [624/625]  eta: 0:00:01  lr: 0.001888  min_lr: 0.001888  loss: 3.0013 (3.0017)  class_acc: 0.5273 (0.5258)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2426 (2.6582)  time: 0.7014  data: 0.0012  max mem: 2905
Epoch: [164] Total time: 0:19:30 (1.8725 s / it)
Averaged stats: lr: 0.001888  min_lr: 0.001888  loss: 3.0013 (3.0016)  class_acc: 0.5273 (0.5250)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2426 (2.6582)
Test:  [ 0/50]  eta: 0:10:39  loss: 2.9423 (2.9423)  acc1: 40.0000 (40.0000)  acc5: 64.0000 (64.0000)  time: 12.7900  data: 12.7567  max mem: 2905
Test:  [10/50]  eta: 0:01:28  loss: 3.1255 (3.0971)  acc1: 38.4000 (39.8545)  acc5: 64.0000 (62.9818)  time: 2.2071  data: 2.1855  max mem: 2905
Test:  [20/50]  eta: 0:00:53  loss: 3.3436 (3.2860)  acc1: 36.0000 (35.8476)  acc5: 60.0000 (60.6857)  time: 1.2203  data: 1.2005  max mem: 2905
Test:  [30/50]  eta: 0:00:28  loss: 3.4506 (3.2799)  acc1: 32.0000 (35.9226)  acc5: 58.4000 (60.4129)  time: 0.9896  data: 0.9700  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 3.3231 (3.3201)  acc1: 34.4000 (35.4732)  acc5: 58.4000 (59.3366)  time: 0.4992  data: 0.4803  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 3.3841 (3.3178)  acc1: 32.8000 (35.5360)  acc5: 56.8000 (59.4880)  time: 0.4767  data: 0.4587  max mem: 2905
Test: Total time: 0:00:47 (0.9560 s / it)
* Acc@1 35.890 Acc@5 60.088 loss 3.286
Accuracy of the model on the 50000 test images: 35.9%
Max accuracy: 50.54%
Epoch: [165]  [  0/625]  eta: 3:33:41  lr: 0.001888  min_lr: 0.001888  loss: 2.9401 (2.9401)  class_acc: 0.5195 (0.5195)  weight_decay: 0.0500 (0.0500)  time: 20.5142  data: 17.2189  max mem: 2905
Epoch: [165]  [200/625]  eta: 0:13:29  lr: 0.001881  min_lr: 0.001881  loss: 3.0212 (2.9828)  class_acc: 0.5234 (0.5307)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7025 (3.0605)  time: 1.6481  data: 0.0008  max mem: 2905
Epoch: [165]  [400/625]  eta: 0:07:02  lr: 0.001874  min_lr: 0.001874  loss: 2.9921 (2.9909)  class_acc: 0.5156 (0.5281)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2314 (2.8759)  time: 1.7111  data: 0.0006  max mem: 2905
Epoch: [165]  [600/625]  eta: 0:00:47  lr: 0.001867  min_lr: 0.001867  loss: 3.0458 (2.9946)  class_acc: 0.5156 (0.5272)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.9788 (inf)  time: 1.9794  data: 0.0009  max mem: 2905
Epoch: [165]  [624/625]  eta: 0:00:01  lr: 0.001866  min_lr: 0.001866  loss: 3.0176 (2.9969)  class_acc: 0.5039 (0.5267)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5893 (inf)  time: 0.6704  data: 0.0013  max mem: 2905
Epoch: [165] Total time: 0:19:24 (1.8633 s / it)
Averaged stats: lr: 0.001866  min_lr: 0.001866  loss: 3.0176 (2.9999)  class_acc: 0.5039 (0.5258)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5893 (inf)
Test:  [ 0/50]  eta: 0:10:24  loss: 3.0633 (3.0633)  acc1: 35.2000 (35.2000)  acc5: 62.4000 (62.4000)  time: 12.4867  data: 12.4549  max mem: 2905
Test:  [10/50]  eta: 0:01:25  loss: 2.9873 (2.9387)  acc1: 40.8000 (40.6545)  acc5: 64.8000 (64.8000)  time: 2.1271  data: 2.1046  max mem: 2905
Test:  [20/50]  eta: 0:00:50  loss: 2.9873 (3.0398)  acc1: 39.2000 (38.9333)  acc5: 64.8000 (63.8857)  time: 1.1543  data: 1.1337  max mem: 2905
Test:  [30/50]  eta: 0:00:28  loss: 3.1121 (3.0090)  acc1: 39.2000 (39.8710)  acc5: 64.0000 (64.6194)  time: 1.0377  data: 1.0176  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 3.1345 (3.1053)  acc1: 39.2000 (38.2049)  acc5: 61.6000 (63.1220)  time: 0.5805  data: 0.5603  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 3.1253 (3.0828)  acc1: 36.8000 (38.6720)  acc5: 63.2000 (63.5680)  time: 0.4977  data: 0.4786  max mem: 2905
Test: Total time: 0:00:48 (0.9611 s / it)
* Acc@1 39.370 Acc@5 63.908 loss 3.041
Accuracy of the model on the 50000 test images: 39.4%
Max accuracy: 50.54%
Epoch: [166]  [  0/625]  eta: 3:32:05  lr: 0.001866  min_lr: 0.001866  loss: 2.9962 (2.9962)  class_acc: 0.5469 (0.5469)  weight_decay: 0.0500 (0.0500)  time: 20.3603  data: 20.2128  max mem: 2905
Epoch: [166]  [200/625]  eta: 0:13:55  lr: 0.001859  min_lr: 0.001859  loss: 2.9943 (2.9773)  class_acc: 0.5352 (0.5296)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4041 (2.7321)  time: 1.8612  data: 0.0006  max mem: 2905
Epoch: [166]  [400/625]  eta: 0:07:08  lr: 0.001852  min_lr: 0.001852  loss: 2.9390 (2.9844)  class_acc: 0.5234 (0.5279)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9155 (2.6368)  time: 1.8199  data: 0.0006  max mem: 2905
Epoch: [166]  [600/625]  eta: 0:00:47  lr: 0.001844  min_lr: 0.001844  loss: 2.9537 (2.9877)  class_acc: 0.5273 (0.5272)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.6367 (2.6694)  time: 1.9442  data: 0.0006  max mem: 2905
Epoch: [166]  [624/625]  eta: 0:00:01  lr: 0.001844  min_lr: 0.001844  loss: 2.9819 (2.9888)  class_acc: 0.5195 (0.5271)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8814 (2.6580)  time: 1.1992  data: 0.0013  max mem: 2905
Epoch: [166] Total time: 0:19:33 (1.8768 s / it)
Averaged stats: lr: 0.001844  min_lr: 0.001844  loss: 2.9819 (2.9984)  class_acc: 0.5195 (0.5262)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8814 (2.6580)
Test:  [ 0/50]  eta: 0:10:06  loss: 3.1573 (3.1573)  acc1: 39.2000 (39.2000)  acc5: 63.2000 (63.2000)  time: 12.1281  data: 12.0890  max mem: 2905
Test:  [10/50]  eta: 0:01:22  loss: 3.3033 (3.2254)  acc1: 38.4000 (36.8000)  acc5: 59.2000 (60.0727)  time: 2.0505  data: 2.0303  max mem: 2905
Test:  [20/50]  eta: 0:00:48  loss: 3.4230 (3.3467)  acc1: 36.0000 (34.8571)  acc5: 56.0000 (58.6667)  time: 1.1077  data: 1.0894  max mem: 2905
Test:  [30/50]  eta: 0:00:29  loss: 3.4230 (3.3456)  acc1: 36.0000 (35.4065)  acc5: 56.0000 (58.6839)  time: 1.1326  data: 1.1125  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 3.4131 (3.3856)  acc1: 33.6000 (35.0049)  acc5: 57.6000 (58.3415)  time: 0.7933  data: 0.7729  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 3.4591 (3.4062)  acc1: 32.8000 (34.6400)  acc5: 57.6000 (57.9840)  time: 0.7385  data: 0.7203  max mem: 2905
Test: Total time: 0:00:50 (1.0188 s / it)
* Acc@1 35.154 Acc@5 59.214 loss 3.369
Accuracy of the model on the 50000 test images: 35.2%
Max accuracy: 50.54%
Epoch: [167]  [  0/625]  eta: 4:44:39  lr: 0.001844  min_lr: 0.001844  loss: 3.0249 (3.0249)  class_acc: 0.5195 (0.5195)  weight_decay: 0.0500 (0.0500)  time: 27.3280  data: 13.1028  max mem: 2905
Epoch: [167]  [200/625]  eta: 0:14:16  lr: 0.001836  min_lr: 0.001836  loss: 2.9187 (2.9747)  class_acc: 0.5312 (0.5336)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6803 (2.3803)  time: 1.9411  data: 0.0097  max mem: 2905
Epoch: [167]  [400/625]  eta: 0:07:16  lr: 0.001829  min_lr: 0.001829  loss: 2.9569 (2.9790)  class_acc: 0.5391 (0.5320)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.5310 (2.6824)  time: 1.8691  data: 0.0007  max mem: 2905
Epoch: [167]  [600/625]  eta: 0:00:48  lr: 0.001822  min_lr: 0.001822  loss: 3.0398 (2.9900)  class_acc: 0.5195 (0.5296)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1429 (2.6636)  time: 1.8791  data: 0.0007  max mem: 2905
Epoch: [167]  [624/625]  eta: 0:00:01  lr: 0.001821  min_lr: 0.001821  loss: 3.0024 (2.9914)  class_acc: 0.5117 (0.5293)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2141 (2.6628)  time: 0.3843  data: 0.0014  max mem: 2905
Epoch: [167] Total time: 0:19:51 (1.9072 s / it)
Averaged stats: lr: 0.001821  min_lr: 0.001821  loss: 3.0024 (2.9929)  class_acc: 0.5117 (0.5279)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2141 (2.6628)
Test:  [ 0/50]  eta: 0:11:17  loss: 2.2886 (2.2886)  acc1: 51.2000 (51.2000)  acc5: 75.2000 (75.2000)  time: 13.5560  data: 13.5230  max mem: 2905
Test:  [10/50]  eta: 0:01:31  loss: 2.6449 (2.5244)  acc1: 44.8000 (46.9091)  acc5: 72.0000 (71.8545)  time: 2.2959  data: 2.2761  max mem: 2905
Test:  [20/50]  eta: 0:00:54  loss: 2.7329 (2.7073)  acc1: 42.4000 (43.5810)  acc5: 68.0000 (68.8762)  time: 1.2291  data: 1.2097  max mem: 2905
Test:  [30/50]  eta: 0:00:30  loss: 2.8250 (2.7163)  acc1: 42.4000 (43.8452)  acc5: 66.4000 (68.9032)  time: 1.0924  data: 1.0724  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 2.6335 (2.7280)  acc1: 44.0000 (43.7073)  acc5: 68.0000 (68.8390)  time: 0.6191  data: 0.5998  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.7958 (2.7493)  acc1: 41.6000 (43.2000)  acc5: 67.2000 (68.4640)  time: 0.5082  data: 0.4895  max mem: 2905
Test: Total time: 0:00:51 (1.0294 s / it)
* Acc@1 43.868 Acc@5 68.928 loss 2.713
Accuracy of the model on the 50000 test images: 43.9%
Max accuracy: 50.54%
Epoch: [168]  [  0/625]  eta: 3:39:52  lr: 0.001821  min_lr: 0.001821  loss: 2.8138 (2.8138)  class_acc: 0.5820 (0.5820)  weight_decay: 0.0500 (0.0500)  time: 21.1080  data: 20.8204  max mem: 2905
Epoch: [168]  [200/625]  eta: 0:13:50  lr: 0.001814  min_lr: 0.001814  loss: 2.9459 (2.9694)  class_acc: 0.5391 (0.5325)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.6658 (2.4502)  time: 1.7739  data: 0.0009  max mem: 2905
Epoch: [168]  [400/625]  eta: 0:07:19  lr: 0.001807  min_lr: 0.001807  loss: 2.9447 (2.9821)  class_acc: 0.5352 (0.5300)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2311 (2.5463)  time: 1.9811  data: 0.0007  max mem: 2905
Epoch: [168]  [600/625]  eta: 0:00:49  lr: 0.001800  min_lr: 0.001800  loss: 2.9581 (2.9858)  class_acc: 0.5312 (0.5300)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7950 (2.6243)  time: 1.9608  data: 0.0007  max mem: 2905
Epoch: [168]  [624/625]  eta: 0:00:01  lr: 0.001799  min_lr: 0.001799  loss: 3.0626 (2.9881)  class_acc: 0.5078 (0.5294)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1495 (2.6073)  time: 0.6393  data: 0.0017  max mem: 2905
Epoch: [168] Total time: 0:20:25 (1.9602 s / it)
Averaged stats: lr: 0.001799  min_lr: 0.001799  loss: 3.0626 (2.9926)  class_acc: 0.5078 (0.5281)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1495 (2.6073)
Test:  [ 0/50]  eta: 0:10:57  loss: 3.2454 (3.2454)  acc1: 34.4000 (34.4000)  acc5: 66.4000 (66.4000)  time: 13.1534  data: 13.1248  max mem: 2905
Test:  [10/50]  eta: 0:01:29  loss: 3.2339 (3.1913)  acc1: 37.6000 (38.2545)  acc5: 62.4000 (60.5818)  time: 2.2336  data: 2.2109  max mem: 2905
Test:  [20/50]  eta: 0:00:52  loss: 3.3253 (3.3924)  acc1: 35.2000 (35.3143)  acc5: 57.6000 (58.5905)  time: 1.1850  data: 1.1644  max mem: 2905
Test:  [30/50]  eta: 0:00:31  loss: 3.4640 (3.3858)  acc1: 35.2000 (35.4839)  acc5: 56.8000 (58.8903)  time: 1.1765  data: 1.1562  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 3.2923 (3.3466)  acc1: 36.0000 (35.7659)  acc5: 59.2000 (59.2976)  time: 0.7691  data: 0.7467  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 3.1683 (3.3408)  acc1: 36.0000 (35.6480)  acc5: 60.0000 (59.4720)  time: 0.7373  data: 0.7171  max mem: 2905
Test: Total time: 0:00:52 (1.0576 s / it)
* Acc@1 36.334 Acc@5 59.828 loss 3.298
Accuracy of the model on the 50000 test images: 36.3%
Max accuracy: 50.54%
Epoch: [169]  [  0/625]  eta: 3:39:41  lr: 0.001799  min_lr: 0.001799  loss: 2.9010 (2.9010)  class_acc: 0.5547 (0.5547)  weight_decay: 0.0500 (0.0500)  time: 21.0897  data: 20.2607  max mem: 2905
Epoch: [169]  [200/625]  eta: 0:14:20  lr: 0.001792  min_lr: 0.001792  loss: 2.9941 (2.9758)  class_acc: 0.5195 (0.5320)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.8153 (2.4579)  time: 2.0700  data: 0.0294  max mem: 2905
Epoch: [169]  [400/625]  eta: 0:07:27  lr: 0.001785  min_lr: 0.001785  loss: 2.9884 (2.9861)  class_acc: 0.5195 (0.5298)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0978 (2.3877)  time: 1.8974  data: 1.5291  max mem: 2905
Epoch: [169]  [600/625]  eta: 0:00:49  lr: 0.001777  min_lr: 0.001777  loss: 3.0000 (2.9893)  class_acc: 0.5117 (0.5283)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9412 (2.4352)  time: 1.8642  data: 1.4349  max mem: 2905
Epoch: [169]  [624/625]  eta: 0:00:01  lr: 0.001777  min_lr: 0.001777  loss: 2.9846 (2.9905)  class_acc: 0.5391 (0.5283)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9362 (2.4144)  time: 0.9707  data: 0.6309  max mem: 2905
Epoch: [169] Total time: 0:20:01 (1.9231 s / it)
Averaged stats: lr: 0.001777  min_lr: 0.001777  loss: 2.9846 (2.9867)  class_acc: 0.5391 (0.5292)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9362 (2.4144)
Test:  [ 0/50]  eta: 0:10:32  loss: 3.3138 (3.3138)  acc1: 31.2000 (31.2000)  acc5: 59.2000 (59.2000)  time: 12.6456  data: 12.6194  max mem: 2905
Test:  [10/50]  eta: 0:01:30  loss: 2.7917 (2.9317)  acc1: 41.6000 (40.9455)  acc5: 66.4000 (64.6545)  time: 2.2689  data: 2.2476  max mem: 2905
Test:  [20/50]  eta: 0:00:53  loss: 2.9900 (3.0767)  acc1: 38.4000 (39.2381)  acc5: 63.2000 (62.4762)  time: 1.2369  data: 1.2167  max mem: 2905
Test:  [30/50]  eta: 0:00:28  loss: 3.1523 (3.0585)  acc1: 37.6000 (39.4323)  acc5: 63.2000 (62.9677)  time: 0.9834  data: 0.9627  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 3.1308 (3.0979)  acc1: 38.4000 (38.8878)  acc5: 61.6000 (62.5561)  time: 0.5155  data: 0.4930  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 3.1458 (3.1031)  acc1: 36.0000 (38.6080)  acc5: 61.6000 (62.8160)  time: 0.4367  data: 0.4162  max mem: 2905
Test: Total time: 0:00:48 (0.9731 s / it)
* Acc@1 39.048 Acc@5 63.846 loss 3.065
Accuracy of the model on the 50000 test images: 39.0%
Max accuracy: 50.54%
Epoch: [170]  [  0/625]  eta: 3:21:06  lr: 0.001777  min_lr: 0.001777  loss: 2.9403 (2.9403)  class_acc: 0.5625 (0.5625)  weight_decay: 0.0500 (0.0500)  time: 19.3067  data: 19.1782  max mem: 2905
Epoch: [170]  [200/625]  eta: 0:13:46  lr: 0.001769  min_lr: 0.001769  loss: 3.0203 (2.9719)  class_acc: 0.5117 (0.5329)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3207 (2.5694)  time: 1.8361  data: 0.0007  max mem: 2905
Epoch: [170]  [400/625]  eta: 0:07:04  lr: 0.001762  min_lr: 0.001762  loss: 2.9403 (2.9849)  class_acc: 0.5391 (0.5309)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1293 (2.8965)  time: 1.8819  data: 0.0007  max mem: 2905
Epoch: [170]  [600/625]  eta: 0:00:46  lr: 0.001755  min_lr: 0.001755  loss: 3.0265 (2.9884)  class_acc: 0.5039 (0.5291)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9512 (2.7736)  time: 1.8766  data: 0.0681  max mem: 2905
Epoch: [170]  [624/625]  eta: 0:00:01  lr: 0.001754  min_lr: 0.001754  loss: 2.9997 (2.9893)  class_acc: 0.5273 (0.5290)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4963 (2.7813)  time: 0.6587  data: 0.0020  max mem: 2905
Epoch: [170] Total time: 0:19:25 (1.8642 s / it)
Averaged stats: lr: 0.001754  min_lr: 0.001754  loss: 2.9997 (2.9856)  class_acc: 0.5273 (0.5297)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4963 (2.7813)
Test:  [ 0/50]  eta: 0:10:26  loss: 3.2164 (3.2164)  acc1: 40.0000 (40.0000)  acc5: 64.8000 (64.8000)  time: 12.5359  data: 12.5076  max mem: 2905
Test:  [10/50]  eta: 0:01:27  loss: 3.0124 (2.9835)  acc1: 41.6000 (41.6727)  acc5: 65.6000 (65.8182)  time: 2.1855  data: 2.1657  max mem: 2905
Test:  [20/50]  eta: 0:00:53  loss: 3.0124 (3.0668)  acc1: 39.2000 (39.6952)  acc5: 65.6000 (64.5333)  time: 1.2393  data: 1.2196  max mem: 2905
Test:  [30/50]  eta: 0:00:31  loss: 3.1176 (3.0541)  acc1: 39.2000 (39.6903)  acc5: 62.4000 (64.1806)  time: 1.2716  data: 1.2511  max mem: 2905
Test:  [40/50]  eta: 0:00:13  loss: 2.9376 (3.0710)  acc1: 39.2000 (39.5902)  acc5: 62.4000 (64.1561)  time: 0.8478  data: 0.8267  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.8790 (3.0384)  acc1: 40.0000 (39.8720)  acc5: 67.2000 (64.4320)  time: 0.8235  data: 0.8031  max mem: 2905
Test: Total time: 0:00:54 (1.0928 s / it)
* Acc@1 40.450 Acc@5 65.534 loss 2.986
Accuracy of the model on the 50000 test images: 40.5%
Max accuracy: 50.54%
Epoch: [171]  [  0/625]  eta: 3:38:26  lr: 0.001754  min_lr: 0.001754  loss: 3.0879 (3.0879)  class_acc: 0.4844 (0.4844)  weight_decay: 0.0500 (0.0500)  time: 20.9705  data: 18.2474  max mem: 2905
Epoch: [171]  [200/625]  eta: 0:14:30  lr: 0.001747  min_lr: 0.001747  loss: 2.9826 (2.9559)  class_acc: 0.5312 (0.5364)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8893 (2.7455)  time: 1.9763  data: 0.4276  max mem: 2905
Epoch: [171]  [400/625]  eta: 0:07:22  lr: 0.001740  min_lr: 0.001740  loss: 2.9612 (2.9695)  class_acc: 0.5312 (0.5340)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2606 (2.7927)  time: 1.8029  data: 0.1976  max mem: 2905
Epoch: [171]  [600/625]  eta: 0:00:48  lr: 0.001733  min_lr: 0.001733  loss: 2.9750 (2.9818)  class_acc: 0.5352 (0.5304)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6586 (2.7298)  time: 1.9153  data: 0.0194  max mem: 2905
Epoch: [171]  [624/625]  eta: 0:00:01  lr: 0.001732  min_lr: 0.001732  loss: 2.9626 (2.9816)  class_acc: 0.5352 (0.5304)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7193 (2.6867)  time: 0.9668  data: 0.0252  max mem: 2905
Epoch: [171] Total time: 0:19:52 (1.9083 s / it)
Averaged stats: lr: 0.001732  min_lr: 0.001732  loss: 2.9626 (2.9838)  class_acc: 0.5352 (0.5299)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7193 (2.6867)
Test:  [ 0/50]  eta: 0:09:43  loss: 3.2135 (3.2135)  acc1: 38.4000 (38.4000)  acc5: 63.2000 (63.2000)  time: 11.6679  data: 11.6320  max mem: 2905
Test:  [10/50]  eta: 0:01:17  loss: 3.1183 (3.0763)  acc1: 38.4000 (39.1273)  acc5: 63.2000 (62.4727)  time: 1.9262  data: 1.9045  max mem: 2905
Test:  [20/50]  eta: 0:00:43  loss: 3.1183 (3.1506)  acc1: 37.6000 (37.6762)  acc5: 63.2000 (62.2476)  time: 0.9504  data: 0.9296  max mem: 2905
Test:  [30/50]  eta: 0:00:25  loss: 3.1413 (3.1643)  acc1: 35.2000 (37.0323)  acc5: 62.4000 (62.1419)  time: 0.9341  data: 0.9131  max mem: 2905
Test:  [40/50]  eta: 0:00:10  loss: 3.2040 (3.1711)  acc1: 36.8000 (37.3659)  acc5: 62.4000 (62.3415)  time: 0.6586  data: 0.6386  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 3.2941 (3.1729)  acc1: 38.4000 (37.6320)  acc5: 62.4000 (62.0000)  time: 0.4853  data: 0.4672  max mem: 2905
Test: Total time: 0:00:44 (0.8985 s / it)
* Acc@1 37.836 Acc@5 62.512 loss 3.137
Accuracy of the model on the 50000 test images: 37.8%
Max accuracy: 50.54%
Epoch: [172]  [  0/625]  eta: 3:44:31  lr: 0.001732  min_lr: 0.001732  loss: 2.9385 (2.9385)  class_acc: 0.5820 (0.5820)  weight_decay: 0.0500 (0.0500)  time: 21.5539  data: 18.0907  max mem: 2905
Epoch: [172]  [200/625]  eta: 0:14:05  lr: 0.001725  min_lr: 0.001725  loss: 2.9877 (2.9806)  class_acc: 0.5195 (0.5315)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1970 (2.5075)  time: 1.8672  data: 0.5351  max mem: 2905
Epoch: [172]  [400/625]  eta: 0:07:16  lr: 0.001718  min_lr: 0.001718  loss: 2.9649 (2.9821)  class_acc: 0.5273 (0.5293)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0467 (inf)  time: 1.9091  data: 0.1682  max mem: 2905
Epoch: [172]  [600/625]  eta: 0:00:47  lr: 0.001711  min_lr: 0.001711  loss: 3.0100 (2.9793)  class_acc: 0.5234 (0.5312)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0628 (inf)  time: 1.9168  data: 0.0007  max mem: 2905
Epoch: [172]  [624/625]  eta: 0:00:01  lr: 0.001710  min_lr: 0.001710  loss: 2.9436 (2.9790)  class_acc: 0.5312 (0.5312)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.1261 (inf)  time: 0.5992  data: 0.0015  max mem: 2905
Epoch: [172] Total time: 0:19:46 (1.8983 s / it)
Averaged stats: lr: 0.001710  min_lr: 0.001710  loss: 2.9436 (2.9805)  class_acc: 0.5312 (0.5308)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.1261 (inf)
Test:  [ 0/50]  eta: 0:11:09  loss: 3.2755 (3.2755)  acc1: 32.8000 (32.8000)  acc5: 61.6000 (61.6000)  time: 13.3831  data: 13.3492  max mem: 2905
Test:  [10/50]  eta: 0:01:29  loss: 2.8431 (2.9376)  acc1: 40.8000 (40.3636)  acc5: 67.2000 (65.3091)  time: 2.2493  data: 2.2292  max mem: 2905
Test:  [20/50]  eta: 0:00:53  loss: 3.0852 (3.1209)  acc1: 36.8000 (37.0667)  acc5: 62.4000 (62.8952)  time: 1.1907  data: 1.1717  max mem: 2905
Test:  [30/50]  eta: 0:00:30  loss: 3.1739 (3.1121)  acc1: 36.8000 (38.0129)  acc5: 60.8000 (62.8387)  time: 1.1428  data: 1.1229  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 3.2810 (3.1978)  acc1: 35.2000 (37.4049)  acc5: 60.8000 (61.8732)  time: 0.7285  data: 0.7089  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 3.2611 (3.1785)  acc1: 35.2000 (37.6320)  acc5: 60.8000 (62.1280)  time: 0.7267  data: 0.7088  max mem: 2905
Test: Total time: 0:00:52 (1.0426 s / it)
* Acc@1 38.084 Acc@5 62.774 loss 3.138
Accuracy of the model on the 50000 test images: 38.1%
Max accuracy: 50.54%
Epoch: [173]  [  0/625]  eta: 3:56:53  lr: 0.001710  min_lr: 0.001710  loss: 2.8024 (2.8024)  class_acc: 0.5430 (0.5430)  weight_decay: 0.0500 (0.0500)  time: 22.7415  data: 20.1273  max mem: 2905
Epoch: [173]  [200/625]  eta: 0:14:24  lr: 0.001703  min_lr: 0.001703  loss: 2.9312 (2.9495)  class_acc: 0.5352 (0.5378)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0731 (2.6881)  time: 1.8458  data: 0.0161  max mem: 2905
Epoch: [173]  [400/625]  eta: 0:07:20  lr: 0.001696  min_lr: 0.001696  loss: 3.0127 (2.9510)  class_acc: 0.5156 (0.5372)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1637 (2.5861)  time: 1.8729  data: 0.0007  max mem: 2905
Epoch: [173]  [600/625]  eta: 0:00:48  lr: 0.001689  min_lr: 0.001689  loss: 2.9601 (2.9667)  class_acc: 0.5195 (0.5341)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4811 (2.7430)  time: 2.0259  data: 0.0015  max mem: 2905
Epoch: [173]  [624/625]  eta: 0:00:01  lr: 0.001688  min_lr: 0.001688  loss: 2.9370 (2.9675)  class_acc: 0.5195 (0.5337)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4555 (2.7457)  time: 0.5736  data: 0.0013  max mem: 2905
Epoch: [173] Total time: 0:20:08 (1.9336 s / it)
Averaged stats: lr: 0.001688  min_lr: 0.001688  loss: 2.9370 (2.9746)  class_acc: 0.5195 (0.5317)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4555 (2.7457)
Test:  [ 0/50]  eta: 0:09:08  loss: 2.6666 (2.6666)  acc1: 40.0000 (40.0000)  acc5: 69.6000 (69.6000)  time: 10.9764  data: 10.9503  max mem: 2905
Test:  [10/50]  eta: 0:01:08  loss: 2.7626 (2.8494)  acc1: 40.8000 (40.8000)  acc5: 66.4000 (65.6000)  time: 1.7093  data: 1.6876  max mem: 2905
Test:  [20/50]  eta: 0:00:38  loss: 2.9249 (2.9672)  acc1: 40.0000 (39.1619)  acc5: 64.8000 (64.2667)  time: 0.7943  data: 0.7713  max mem: 2905
Test:  [30/50]  eta: 0:00:24  loss: 3.1269 (3.0094)  acc1: 36.8000 (39.0710)  acc5: 61.6000 (63.1484)  time: 0.9874  data: 0.9659  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 3.1472 (3.0558)  acc1: 36.8000 (38.8878)  acc5: 60.8000 (62.4195)  time: 1.0054  data: 0.9864  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 3.1000 (3.0549)  acc1: 36.8000 (39.1040)  acc5: 62.4000 (62.7840)  time: 0.5792  data: 0.5598  max mem: 2905
Test: Total time: 0:00:50 (1.0061 s / it)
* Acc@1 39.206 Acc@5 63.740 loss 3.025
Accuracy of the model on the 50000 test images: 39.2%
Max accuracy: 50.54%
Epoch: [174]  [  0/625]  eta: 3:39:08  lr: 0.001688  min_lr: 0.001688  loss: 3.0976 (3.0976)  class_acc: 0.5117 (0.5117)  weight_decay: 0.0500 (0.0500)  time: 21.0382  data: 19.9410  max mem: 2905
Epoch: [174]  [200/625]  eta: 0:14:11  lr: 0.001681  min_lr: 0.001681  loss: 2.9163 (2.9552)  class_acc: 0.5312 (0.5368)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7128 (2.7697)  time: 1.8328  data: 0.0419  max mem: 2905
Epoch: [174]  [400/625]  eta: 0:07:26  lr: 0.001674  min_lr: 0.001674  loss: 2.9233 (2.9677)  class_acc: 0.5352 (0.5334)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4044 (2.6798)  time: 1.9221  data: 0.0007  max mem: 2905
Epoch: [174]  [600/625]  eta: 0:00:49  lr: 0.001666  min_lr: 0.001666  loss: 2.9966 (2.9767)  class_acc: 0.5273 (0.5307)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3354 (2.6386)  time: 2.0869  data: 0.0007  max mem: 2905
Epoch: [174]  [624/625]  eta: 0:00:01  lr: 0.001666  min_lr: 0.001666  loss: 3.0110 (2.9780)  class_acc: 0.5273 (0.5308)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3033 (2.6561)  time: 0.7380  data: 0.0014  max mem: 2905
Epoch: [174] Total time: 0:20:08 (1.9341 s / it)
Averaged stats: lr: 0.001666  min_lr: 0.001666  loss: 3.0110 (2.9749)  class_acc: 0.5273 (0.5315)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3033 (2.6561)
Test:  [ 0/50]  eta: 0:10:01  loss: 3.0384 (3.0384)  acc1: 36.8000 (36.8000)  acc5: 59.2000 (59.2000)  time: 12.0253  data: 11.9985  max mem: 2905
Test:  [10/50]  eta: 0:01:18  loss: 3.0384 (2.9753)  acc1: 37.6000 (40.3636)  acc5: 62.4000 (63.2000)  time: 1.9516  data: 1.9317  max mem: 2905
Test:  [20/50]  eta: 0:00:43  loss: 3.0979 (3.1068)  acc1: 37.6000 (37.5619)  acc5: 61.6000 (61.8667)  time: 0.9101  data: 0.8908  max mem: 2905
Test:  [30/50]  eta: 0:00:24  loss: 3.0826 (3.0781)  acc1: 37.6000 (38.4000)  acc5: 62.4000 (62.6323)  time: 0.8580  data: 0.8374  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 3.1173 (3.1519)  acc1: 40.0000 (38.0878)  acc5: 60.8000 (61.7366)  time: 0.8465  data: 0.8265  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 3.1689 (3.1511)  acc1: 32.8000 (37.4560)  acc5: 59.2000 (61.7920)  time: 0.7235  data: 0.7052  max mem: 2905
Test: Total time: 0:00:51 (1.0305 s / it)
* Acc@1 37.862 Acc@5 62.306 loss 3.122
Accuracy of the model on the 50000 test images: 37.9%
Max accuracy: 50.54%
Epoch: [175]  [  0/625]  eta: 3:45:41  lr: 0.001666  min_lr: 0.001666  loss: 2.9276 (2.9276)  class_acc: 0.5352 (0.5352)  weight_decay: 0.0500 (0.0500)  time: 21.6667  data: 18.7025  max mem: 2905
Epoch: [175]  [200/625]  eta: 0:14:08  lr: 0.001658  min_lr: 0.001658  loss: 2.9273 (2.9511)  class_acc: 0.5352 (0.5380)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1802 (inf)  time: 1.8904  data: 0.0972  max mem: 2905
Epoch: [175]  [400/625]  eta: 0:07:18  lr: 0.001651  min_lr: 0.001651  loss: 2.9357 (2.9673)  class_acc: 0.5312 (0.5349)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2905 (inf)  time: 1.7468  data: 0.0189  max mem: 2905
Epoch: [175]  [600/625]  eta: 0:00:49  lr: 0.001644  min_lr: 0.001644  loss: 2.9760 (2.9734)  class_acc: 0.5117 (0.5335)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2182 (inf)  time: 1.7601  data: 0.0112  max mem: 2905
Epoch: [175]  [624/625]  eta: 0:00:01  lr: 0.001644  min_lr: 0.001644  loss: 2.9708 (2.9740)  class_acc: 0.5312 (0.5333)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9865 (inf)  time: 0.8576  data: 0.0134  max mem: 2905
Epoch: [175] Total time: 0:19:59 (1.9195 s / it)
Averaged stats: lr: 0.001644  min_lr: 0.001644  loss: 2.9708 (2.9721)  class_acc: 0.5312 (0.5322)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9865 (inf)
Test:  [ 0/50]  eta: 0:10:03  loss: 3.7714 (3.7714)  acc1: 37.6000 (37.6000)  acc5: 56.0000 (56.0000)  time: 12.0751  data: 12.0456  max mem: 2905
Test:  [10/50]  eta: 0:01:21  loss: 3.8978 (3.9093)  acc1: 32.8000 (29.9636)  acc5: 51.2000 (50.9818)  time: 2.0429  data: 2.0232  max mem: 2905
Test:  [20/50]  eta: 0:00:46  loss: 3.9049 (3.9608)  acc1: 26.4000 (27.8476)  acc5: 48.8000 (50.5905)  time: 1.0085  data: 0.9883  max mem: 2905
Test:  [30/50]  eta: 0:00:25  loss: 4.0368 (3.9828)  acc1: 25.6000 (27.5871)  acc5: 48.8000 (50.2710)  time: 0.8854  data: 0.8645  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 4.0735 (4.0297)  acc1: 25.6000 (27.0049)  acc5: 49.6000 (50.2049)  time: 0.7169  data: 0.6953  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 4.1491 (4.0414)  acc1: 25.6000 (27.1840)  acc5: 49.6000 (49.9840)  time: 0.6759  data: 0.6550  max mem: 2905
Test: Total time: 0:00:50 (1.0072 s / it)
* Acc@1 27.642 Acc@5 50.460 loss 4.006
Accuracy of the model on the 50000 test images: 27.6%
Max accuracy: 50.54%
Epoch: [176]  [  0/625]  eta: 3:45:17  lr: 0.001643  min_lr: 0.001643  loss: 2.8093 (2.8093)  class_acc: 0.5586 (0.5586)  weight_decay: 0.0500 (0.0500)  time: 21.6274  data: 21.4630  max mem: 2905
Epoch: [176]  [200/625]  eta: 0:14:25  lr: 0.001636  min_lr: 0.001636  loss: 2.9613 (2.9424)  class_acc: 0.5352 (0.5377)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7073 (2.8015)  time: 1.8253  data: 0.0006  max mem: 2905
Epoch: [176]  [400/625]  eta: 0:07:28  lr: 0.001629  min_lr: 0.001629  loss: 2.9726 (2.9564)  class_acc: 0.5312 (0.5350)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1197 (2.8548)  time: 2.0773  data: 0.0007  max mem: 2905
Epoch: [176]  [600/625]  eta: 0:00:49  lr: 0.001622  min_lr: 0.001622  loss: 2.9676 (2.9622)  class_acc: 0.5312 (0.5343)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3538 (2.8196)  time: 1.9686  data: 0.0006  max mem: 2905
Epoch: [176]  [624/625]  eta: 0:00:01  lr: 0.001621  min_lr: 0.001621  loss: 3.0047 (2.9636)  class_acc: 0.5312 (0.5343)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3463 (2.8120)  time: 0.7437  data: 0.0013  max mem: 2905
Epoch: [176] Total time: 0:20:16 (1.9468 s / it)
Averaged stats: lr: 0.001621  min_lr: 0.001621  loss: 3.0047 (2.9677)  class_acc: 0.5312 (0.5332)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3463 (2.8120)
Test:  [ 0/50]  eta: 0:10:48  loss: 3.6266 (3.6266)  acc1: 32.0000 (32.0000)  acc5: 53.6000 (53.6000)  time: 12.9700  data: 12.9368  max mem: 2905
Test:  [10/50]  eta: 0:01:29  loss: 3.4445 (3.3868)  acc1: 33.6000 (34.4000)  acc5: 57.6000 (57.2364)  time: 2.2257  data: 2.2048  max mem: 2905
Test:  [20/50]  eta: 0:00:53  loss: 3.2033 (3.3318)  acc1: 37.6000 (35.3524)  acc5: 58.4000 (58.9714)  time: 1.2245  data: 1.2056  max mem: 2905
Test:  [30/50]  eta: 0:00:32  loss: 3.3299 (3.3366)  acc1: 36.8000 (36.1548)  acc5: 59.2000 (59.2000)  time: 1.2774  data: 1.2586  max mem: 2905
Test:  [40/50]  eta: 0:00:13  loss: 3.3606 (3.3586)  acc1: 36.0000 (36.0390)  acc5: 58.4000 (58.9659)  time: 0.9229  data: 0.9035  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 3.3969 (3.3777)  acc1: 34.4000 (35.8240)  acc5: 58.4000 (58.6560)  time: 0.8259  data: 0.8054  max mem: 2905
Test: Total time: 0:00:56 (1.1291 s / it)
* Acc@1 36.232 Acc@5 59.528 loss 3.320
Accuracy of the model on the 50000 test images: 36.2%
Max accuracy: 50.54%
Epoch: [177]  [  0/625]  eta: 3:33:21  lr: 0.001621  min_lr: 0.001621  loss: 2.8986 (2.8986)  class_acc: 0.5352 (0.5352)  weight_decay: 0.0500 (0.0500)  time: 20.4823  data: 18.8573  max mem: 2905
Epoch: [177]  [200/625]  eta: 0:14:06  lr: 0.001614  min_lr: 0.001614  loss: 2.9350 (2.9588)  class_acc: 0.5469 (0.5372)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3800 (3.0986)  time: 1.7903  data: 0.0164  max mem: 2905
Epoch: [177]  [400/625]  eta: 0:07:24  lr: 0.001607  min_lr: 0.001607  loss: 2.9678 (2.9623)  class_acc: 0.5352 (0.5356)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5470 (3.0213)  time: 2.1067  data: 0.0009  max mem: 2905
Epoch: [177]  [600/625]  eta: 0:00:48  lr: 0.001600  min_lr: 0.001600  loss: 2.9456 (2.9672)  class_acc: 0.5273 (0.5340)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9555 (2.8579)  time: 1.9036  data: 0.0007  max mem: 2905
Epoch: [177]  [624/625]  eta: 0:00:01  lr: 0.001599  min_lr: 0.001599  loss: 2.9981 (2.9696)  class_acc: 0.5273 (0.5336)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2223 (2.8605)  time: 0.7163  data: 0.0027  max mem: 2905
Epoch: [177] Total time: 0:20:07 (1.9326 s / it)
Averaged stats: lr: 0.001599  min_lr: 0.001599  loss: 2.9981 (2.9661)  class_acc: 0.5273 (0.5336)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2223 (2.8605)
Test:  [ 0/50]  eta: 0:09:34  loss: 2.1445 (2.1445)  acc1: 56.8000 (56.8000)  acc5: 78.4000 (78.4000)  time: 11.4892  data: 11.4387  max mem: 2905
Test:  [10/50]  eta: 0:01:15  loss: 2.0917 (2.1205)  acc1: 56.8000 (54.9818)  acc5: 77.6000 (78.3273)  time: 1.8893  data: 1.8615  max mem: 2905
Test:  [20/50]  eta: 0:00:44  loss: 2.2997 (2.3323)  acc1: 48.0000 (50.7048)  acc5: 73.6000 (75.2000)  time: 0.9762  data: 0.9528  max mem: 2905
Test:  [30/50]  eta: 0:00:25  loss: 2.5339 (2.3686)  acc1: 46.4000 (49.9871)  acc5: 72.0000 (74.4774)  time: 0.8995  data: 0.8784  max mem: 2905
Test:  [40/50]  eta: 0:00:10  loss: 2.3273 (2.3865)  acc1: 48.0000 (49.7561)  acc5: 72.8000 (74.2049)  time: 0.5945  data: 0.5745  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 2.3960 (2.3967)  acc1: 48.8000 (49.6960)  acc5: 73.6000 (74.0320)  time: 0.5065  data: 0.4874  max mem: 2905
Test: Total time: 0:00:45 (0.9134 s / it)
* Acc@1 49.656 Acc@5 74.674 loss 2.365
Accuracy of the model on the 50000 test images: 49.7%
Max accuracy: 50.54%
Epoch: [178]  [  0/625]  eta: 3:36:01  lr: 0.001599  min_lr: 0.001599  loss: 3.2124 (3.2124)  class_acc: 0.5352 (0.5352)  weight_decay: 0.0500 (0.0500)  time: 20.7386  data: 20.3402  max mem: 2905
Epoch: [178]  [200/625]  eta: 0:14:44  lr: 0.001592  min_lr: 0.001592  loss: 2.9707 (2.9610)  class_acc: 0.5156 (0.5361)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.0166 (2.7647)  time: 1.9292  data: 0.0094  max mem: 2905
Epoch: [178]  [400/625]  eta: 0:07:27  lr: 0.001585  min_lr: 0.001585  loss: 2.9206 (2.9634)  class_acc: 0.5312 (0.5339)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.9752 (2.8134)  time: 2.0765  data: 0.0007  max mem: 2905
Epoch: [178]  [600/625]  eta: 0:00:49  lr: 0.001578  min_lr: 0.001578  loss: 2.9737 (2.9597)  class_acc: 0.5352 (0.5343)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7499 (2.7650)  time: 2.0173  data: 0.0012  max mem: 2905
Epoch: [178]  [624/625]  eta: 0:00:01  lr: 0.001578  min_lr: 0.001578  loss: 3.0011 (2.9596)  class_acc: 0.5273 (0.5343)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1669 (2.7711)  time: 0.7741  data: 0.0014  max mem: 2905
Epoch: [178] Total time: 0:20:12 (1.9405 s / it)
Averaged stats: lr: 0.001578  min_lr: 0.001578  loss: 3.0011 (2.9628)  class_acc: 0.5273 (0.5338)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1669 (2.7711)
Test:  [ 0/50]  eta: 0:09:33  loss: 3.1583 (3.1583)  acc1: 35.2000 (35.2000)  acc5: 64.0000 (64.0000)  time: 11.4778  data: 11.4509  max mem: 2905
Test:  [10/50]  eta: 0:01:26  loss: 2.6522 (2.6764)  acc1: 44.8000 (43.2000)  acc5: 68.0000 (69.7455)  time: 2.1689  data: 2.1478  max mem: 2905
Test:  [20/50]  eta: 0:00:54  loss: 2.8025 (2.7664)  acc1: 42.4000 (41.8286)  acc5: 67.2000 (68.1143)  time: 1.3218  data: 1.3005  max mem: 2905
Test:  [30/50]  eta: 0:00:31  loss: 2.8040 (2.7718)  acc1: 40.8000 (42.9677)  acc5: 67.2000 (68.3355)  time: 1.2171  data: 1.1953  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 2.8040 (2.8160)  acc1: 40.8000 (42.6537)  acc5: 67.2000 (67.7073)  time: 0.7104  data: 0.6884  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.8978 (2.8076)  acc1: 40.8000 (42.8480)  acc5: 67.2000 (67.6800)  time: 0.6119  data: 0.5909  max mem: 2905
Test: Total time: 0:00:54 (1.0800 s / it)
* Acc@1 43.564 Acc@5 68.272 loss 2.770
Accuracy of the model on the 50000 test images: 43.6%
Max accuracy: 50.54%
Epoch: [179]  [  0/625]  eta: 3:23:31  lr: 0.001577  min_lr: 0.001577  loss: 3.0459 (3.0459)  class_acc: 0.5078 (0.5078)  weight_decay: 0.0500 (0.0500)  time: 19.5389  data: 18.4212  max mem: 2905
Epoch: [179]  [200/625]  eta: 0:15:29  lr: 0.001570  min_lr: 0.001570  loss: 2.9030 (2.9636)  class_acc: 0.5391 (0.5331)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1817 (2.6462)  time: 2.0051  data: 0.0005  max mem: 2905
Epoch: [179]  [400/625]  eta: 0:07:52  lr: 0.001563  min_lr: 0.001563  loss: 2.9813 (2.9597)  class_acc: 0.5234 (0.5350)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.1210 (2.7472)  time: 1.8979  data: 0.0006  max mem: 2905
Epoch: [179]  [600/625]  eta: 0:00:51  lr: 0.001556  min_lr: 0.001556  loss: 2.9118 (2.9608)  class_acc: 0.5312 (0.5347)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1375 (2.7450)  time: 2.0920  data: 0.0006  max mem: 2905
Epoch: [179]  [624/625]  eta: 0:00:02  lr: 0.001556  min_lr: 0.001556  loss: 2.8960 (2.9600)  class_acc: 0.5352 (0.5348)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3293 (2.7545)  time: 1.0082  data: 0.0012  max mem: 2905
Epoch: [179] Total time: 0:20:55 (2.0094 s / it)
Averaged stats: lr: 0.001556  min_lr: 0.001556  loss: 2.8960 (2.9562)  class_acc: 0.5352 (0.5361)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3293 (2.7545)
Test:  [ 0/50]  eta: 0:08:28  loss: 3.0974 (3.0974)  acc1: 37.6000 (37.6000)  acc5: 59.2000 (59.2000)  time: 10.1621  data: 10.1307  max mem: 2905
Test:  [10/50]  eta: 0:01:19  loss: 2.6009 (2.6745)  acc1: 44.0000 (43.4909)  acc5: 69.6000 (68.4364)  time: 1.9976  data: 1.9784  max mem: 2905
Test:  [20/50]  eta: 0:00:47  loss: 2.7106 (2.9242)  acc1: 40.8000 (40.5714)  acc5: 68.0000 (65.4857)  time: 1.1709  data: 1.1520  max mem: 2905
Test:  [30/50]  eta: 0:00:26  loss: 3.0709 (2.9526)  acc1: 40.0000 (40.8000)  acc5: 62.4000 (65.3677)  time: 0.9781  data: 0.9587  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 2.9931 (2.9606)  acc1: 40.8000 (41.0146)  acc5: 64.0000 (65.2098)  time: 0.7432  data: 0.7245  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.9915 (2.9503)  acc1: 40.8000 (40.9600)  acc5: 64.8000 (65.4080)  time: 0.6526  data: 0.6346  max mem: 2905
Test: Total time: 0:00:51 (1.0331 s / it)
* Acc@1 41.638 Acc@5 66.288 loss 2.904
Accuracy of the model on the 50000 test images: 41.6%
Max accuracy: 50.54%
Epoch: [180]  [  0/625]  eta: 3:54:11  lr: 0.001556  min_lr: 0.001556  loss: 3.0688 (3.0688)  class_acc: 0.5273 (0.5273)  weight_decay: 0.0500 (0.0500)  time: 22.4827  data: 19.3503  max mem: 2905
Epoch: [180]  [200/625]  eta: 0:16:56  lr: 0.001549  min_lr: 0.001549  loss: 2.9402 (2.9402)  class_acc: 0.5469 (0.5408)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.8623 (2.6197)  time: 2.0191  data: 0.0006  max mem: 2905
Epoch: [180]  [400/625]  eta: 0:08:07  lr: 0.001542  min_lr: 0.001542  loss: 2.9974 (2.9514)  class_acc: 0.5195 (0.5375)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0796 (2.5966)  time: 1.8534  data: 0.0007  max mem: 2905
Epoch: [180]  [600/625]  eta: 0:00:52  lr: 0.001535  min_lr: 0.001535  loss: 2.9739 (2.9596)  class_acc: 0.5430 (0.5363)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.2341 (2.7636)  time: 1.8660  data: 0.0006  max mem: 2905
Epoch: [180]  [624/625]  eta: 0:00:02  lr: 0.001534  min_lr: 0.001534  loss: 2.9460 (2.9591)  class_acc: 0.5273 (0.5364)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0194 (2.7407)  time: 0.7925  data: 0.0015  max mem: 2905
Epoch: [180] Total time: 0:21:14 (2.0386 s / it)
Averaged stats: lr: 0.001534  min_lr: 0.001534  loss: 2.9460 (2.9575)  class_acc: 0.5273 (0.5356)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0194 (2.7407)
Test:  [ 0/50]  eta: 0:09:47  loss: 3.0501 (3.0501)  acc1: 40.8000 (40.8000)  acc5: 62.4000 (62.4000)  time: 11.7423  data: 11.7172  max mem: 2905
Test:  [10/50]  eta: 0:01:22  loss: 2.5562 (2.6453)  acc1: 42.4000 (44.0727)  acc5: 68.8000 (69.0182)  time: 2.0555  data: 2.0329  max mem: 2905
Test:  [20/50]  eta: 0:00:50  loss: 2.7856 (2.8169)  acc1: 42.4000 (42.2857)  acc5: 68.0000 (67.3524)  time: 1.1817  data: 1.1594  max mem: 2905
Test:  [30/50]  eta: 0:00:30  loss: 2.8759 (2.8285)  acc1: 40.0000 (42.4000)  acc5: 66.4000 (67.5097)  time: 1.1980  data: 1.1764  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 2.8759 (2.8659)  acc1: 40.8000 (42.3024)  acc5: 66.4000 (66.8488)  time: 0.7998  data: 0.7791  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.8480 (2.8717)  acc1: 40.8000 (42.1760)  acc5: 66.4000 (66.7200)  time: 0.6438  data: 0.6228  max mem: 2905
Test: Total time: 0:00:52 (1.0520 s / it)
* Acc@1 42.880 Acc@5 67.184 loss 2.819
Accuracy of the model on the 50000 test images: 42.9%
Max accuracy: 50.54%
Epoch: [181]  [  0/625]  eta: 3:48:01  lr: 0.001534  min_lr: 0.001534  loss: 2.9576 (2.9576)  class_acc: 0.5312 (0.5312)  weight_decay: 0.0500 (0.0500)  time: 21.8904  data: 19.8802  max mem: 2905
Epoch: [181]  [200/625]  eta: 0:13:35  lr: 0.001527  min_lr: 0.001527  loss: 2.9780 (2.9274)  class_acc: 0.5312 (0.5427)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6591 (2.6912)  time: 1.8400  data: 0.0272  max mem: 2905
Epoch: [181]  [400/625]  eta: 0:07:04  lr: 0.001520  min_lr: 0.001520  loss: 3.0450 (2.9455)  class_acc: 0.5156 (0.5370)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.0059 (2.8901)  time: 1.7974  data: 0.0065  max mem: 2905
Epoch: [181]  [600/625]  eta: 0:00:47  lr: 0.001513  min_lr: 0.001513  loss: 2.9584 (2.9560)  class_acc: 0.5352 (0.5359)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7979 (2.8036)  time: 2.0169  data: 0.0010  max mem: 2905
Epoch: [181]  [624/625]  eta: 0:00:01  lr: 0.001512  min_lr: 0.001512  loss: 2.9850 (2.9582)  class_acc: 0.5234 (0.5353)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0999 (2.8138)  time: 0.6392  data: 0.0017  max mem: 2905
Epoch: [181] Total time: 0:19:33 (1.8784 s / it)
Averaged stats: lr: 0.001512  min_lr: 0.001512  loss: 2.9850 (2.9535)  class_acc: 0.5234 (0.5368)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0999 (2.8138)
Test:  [ 0/50]  eta: 0:10:25  loss: 3.1346 (3.1346)  acc1: 40.8000 (40.8000)  acc5: 64.0000 (64.0000)  time: 12.5050  data: 12.4746  max mem: 2905
Test:  [10/50]  eta: 0:01:19  loss: 3.0459 (3.0258)  acc1: 42.4000 (40.7273)  acc5: 64.0000 (64.3636)  time: 1.9912  data: 1.9724  max mem: 2905
Test:  [20/50]  eta: 0:00:46  loss: 3.1529 (3.1968)  acc1: 37.6000 (37.7905)  acc5: 61.6000 (61.7905)  time: 1.0173  data: 0.9992  max mem: 2905
Test:  [30/50]  eta: 0:00:27  loss: 3.1529 (3.1847)  acc1: 37.6000 (37.8065)  acc5: 61.6000 (61.9613)  time: 1.0406  data: 1.0195  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 3.1476 (3.1930)  acc1: 37.6000 (37.5415)  acc5: 60.0000 (61.6976)  time: 0.7097  data: 0.6879  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 3.2742 (3.2207)  acc1: 36.8000 (37.2960)  acc5: 60.0000 (61.4560)  time: 0.5925  data: 0.5717  max mem: 2905
Test: Total time: 0:00:48 (0.9624 s / it)
* Acc@1 37.396 Acc@5 62.212 loss 3.169
Accuracy of the model on the 50000 test images: 37.4%
Max accuracy: 50.54%
Epoch: [182]  [  0/625]  eta: 3:48:26  lr: 0.001512  min_lr: 0.001512  loss: 2.8528 (2.8528)  class_acc: 0.5742 (0.5742)  weight_decay: 0.0500 (0.0500)  time: 21.9312  data: 17.4945  max mem: 2905
Epoch: [182]  [200/625]  eta: 0:14:16  lr: 0.001505  min_lr: 0.001505  loss: 2.8631 (2.9374)  class_acc: 0.5469 (0.5404)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.8294 (3.1070)  time: 1.8812  data: 0.0701  max mem: 2905
Epoch: [182]  [400/625]  eta: 0:07:22  lr: 0.001498  min_lr: 0.001498  loss: 2.9454 (2.9441)  class_acc: 0.5430 (0.5391)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1880 (2.9962)  time: 1.8618  data: 0.0769  max mem: 2905
Epoch: [182]  [600/625]  eta: 0:00:49  lr: 0.001491  min_lr: 0.001491  loss: 2.9813 (2.9474)  class_acc: 0.5391 (0.5387)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5305 (2.8895)  time: 2.0511  data: 0.0008  max mem: 2905
Epoch: [182]  [624/625]  eta: 0:00:01  lr: 0.001490  min_lr: 0.001490  loss: 2.9624 (2.9483)  class_acc: 0.5234 (0.5383)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.9084 (2.9022)  time: 0.9062  data: 0.0014  max mem: 2905
Epoch: [182] Total time: 0:20:07 (1.9318 s / it)
Averaged stats: lr: 0.001490  min_lr: 0.001490  loss: 2.9624 (2.9506)  class_acc: 0.5234 (0.5369)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.9084 (2.9022)
Test:  [ 0/50]  eta: 0:09:59  loss: 2.4990 (2.4990)  acc1: 44.8000 (44.8000)  acc5: 76.8000 (76.8000)  time: 11.9823  data: 11.9593  max mem: 2905
Test:  [10/50]  eta: 0:01:23  loss: 2.4903 (2.5872)  acc1: 44.8000 (46.1091)  acc5: 72.0000 (70.3273)  time: 2.0939  data: 2.0736  max mem: 2905
Test:  [20/50]  eta: 0:00:50  loss: 2.8230 (2.7645)  acc1: 44.0000 (44.3048)  acc5: 68.8000 (68.6476)  time: 1.1701  data: 1.1502  max mem: 2905
Test:  [30/50]  eta: 0:00:30  loss: 2.9380 (2.7739)  acc1: 41.6000 (44.1032)  acc5: 67.2000 (68.4129)  time: 1.1864  data: 1.1676  max mem: 2905
Test:  [40/50]  eta: 0:00:13  loss: 2.9707 (2.8075)  acc1: 40.8000 (43.6878)  acc5: 68.0000 (68.1561)  time: 0.9251  data: 0.9052  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.7791 (2.8076)  acc1: 41.6000 (43.6800)  acc5: 68.0000 (68.0960)  time: 0.8109  data: 0.7904  max mem: 2905
Test: Total time: 0:00:55 (1.1001 s / it)
* Acc@1 43.998 Acc@5 68.302 loss 2.784
Accuracy of the model on the 50000 test images: 44.0%
Max accuracy: 50.54%
Epoch: [183]  [  0/625]  eta: 3:22:52  lr: 0.001490  min_lr: 0.001490  loss: 3.0205 (3.0205)  class_acc: 0.5156 (0.5156)  weight_decay: 0.0500 (0.0500)  time: 19.4760  data: 17.6965  max mem: 2905
Epoch: [183]  [200/625]  eta: 0:14:14  lr: 0.001483  min_lr: 0.001483  loss: 2.9576 (2.9469)  class_acc: 0.5273 (0.5378)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.8342 (2.7885)  time: 1.9214  data: 0.0647  max mem: 2905
Epoch: [183]  [400/625]  eta: 0:07:25  lr: 0.001476  min_lr: 0.001476  loss: 2.9856 (2.9431)  class_acc: 0.5312 (0.5392)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2166 (2.7283)  time: 2.0967  data: 0.0007  max mem: 2905
Epoch: [183]  [600/625]  eta: 0:00:49  lr: 0.001469  min_lr: 0.001469  loss: 2.9512 (2.9428)  class_acc: 0.5352 (0.5397)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1238 (2.6808)  time: 1.9924  data: 0.0008  max mem: 2905
Epoch: [183]  [624/625]  eta: 0:00:01  lr: 0.001469  min_lr: 0.001469  loss: 2.9977 (2.9445)  class_acc: 0.5195 (0.5391)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5755 (2.6919)  time: 0.8826  data: 0.0062  max mem: 2905
Epoch: [183] Total time: 0:20:15 (1.9456 s / it)
Averaged stats: lr: 0.001469  min_lr: 0.001469  loss: 2.9977 (2.9468)  class_acc: 0.5195 (0.5378)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5755 (2.6919)
Test:  [ 0/50]  eta: 0:10:03  loss: 4.2062 (4.2062)  acc1: 27.2000 (27.2000)  acc5: 45.6000 (45.6000)  time: 12.0727  data: 12.0391  max mem: 2905
Test:  [10/50]  eta: 0:01:19  loss: 3.4823 (3.5067)  acc1: 35.2000 (34.8364)  acc5: 57.6000 (56.8000)  time: 1.9913  data: 1.9704  max mem: 2905
Test:  [20/50]  eta: 0:00:46  loss: 3.4944 (3.6496)  acc1: 32.8000 (32.3429)  acc5: 56.0000 (54.7048)  time: 1.0253  data: 1.0060  max mem: 2905
Test:  [30/50]  eta: 0:00:27  loss: 3.5570 (3.6233)  acc1: 30.4000 (32.8516)  acc5: 55.2000 (54.9161)  time: 1.0616  data: 1.0405  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 3.4127 (3.5927)  acc1: 34.4000 (33.2098)  acc5: 56.8000 (55.6293)  time: 0.9404  data: 0.9192  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 3.4998 (3.6264)  acc1: 32.0000 (32.6880)  acc5: 55.2000 (55.3280)  time: 0.8402  data: 0.8208  max mem: 2905
Test: Total time: 0:00:56 (1.1233 s / it)
* Acc@1 32.940 Acc@5 55.742 loss 3.610
Accuracy of the model on the 50000 test images: 32.9%
Max accuracy: 50.54%
Epoch: [184]  [  0/625]  eta: 3:36:10  lr: 0.001469  min_lr: 0.001469  loss: 2.9707 (2.9707)  class_acc: 0.5820 (0.5820)  weight_decay: 0.0500 (0.0500)  time: 20.7529  data: 20.6271  max mem: 2905
Epoch: [184]  [200/625]  eta: 0:14:28  lr: 0.001462  min_lr: 0.001462  loss: 2.9288 (2.9270)  class_acc: 0.5430 (0.5420)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2709 (3.1388)  time: 2.0056  data: 0.0452  max mem: 2905
Epoch: [184]  [400/625]  eta: 0:07:29  lr: 0.001455  min_lr: 0.001455  loss: 2.9200 (2.9405)  class_acc: 0.5391 (0.5378)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3547 (3.0991)  time: 1.8497  data: 0.0008  max mem: 2905
Epoch: [184]  [600/625]  eta: 0:00:49  lr: 0.001448  min_lr: 0.001448  loss: 2.9480 (2.9479)  class_acc: 0.5312 (0.5376)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2036 (2.9126)  time: 2.0344  data: 0.0109  max mem: 2905
Epoch: [184]  [624/625]  eta: 0:00:01  lr: 0.001447  min_lr: 0.001447  loss: 2.9182 (2.9483)  class_acc: 0.5352 (0.5376)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3840 (2.9224)  time: 0.5753  data: 0.0020  max mem: 2905
Epoch: [184] Total time: 0:20:22 (1.9562 s / it)
Averaged stats: lr: 0.001447  min_lr: 0.001447  loss: 2.9182 (2.9463)  class_acc: 0.5352 (0.5379)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3840 (2.9224)
Test:  [ 0/50]  eta: 0:09:24  loss: 2.3522 (2.3522)  acc1: 46.4000 (46.4000)  acc5: 77.6000 (77.6000)  time: 11.2956  data: 11.2701  max mem: 2905
Test:  [10/50]  eta: 0:01:18  loss: 2.4138 (2.4112)  acc1: 50.4000 (49.7455)  acc5: 73.6000 (74.0364)  time: 1.9634  data: 1.9442  max mem: 2905
Test:  [20/50]  eta: 0:00:47  loss: 2.5622 (2.5294)  acc1: 45.6000 (47.2000)  acc5: 72.0000 (72.0762)  time: 1.0864  data: 1.0666  max mem: 2905
Test:  [30/50]  eta: 0:00:27  loss: 2.6801 (2.5933)  acc1: 41.6000 (46.5806)  acc5: 68.8000 (71.2000)  time: 1.0454  data: 1.0250  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 2.6966 (2.6169)  acc1: 44.0000 (46.5561)  acc5: 68.8000 (71.0439)  time: 0.6260  data: 0.6055  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 2.6455 (2.6357)  acc1: 44.0000 (46.3360)  acc5: 71.2000 (70.7520)  time: 0.5272  data: 0.5077  max mem: 2905
Test: Total time: 0:00:46 (0.9321 s / it)
* Acc@1 46.432 Acc@5 71.606 loss 2.580
Accuracy of the model on the 50000 test images: 46.4%
Max accuracy: 50.54%
Epoch: [185]  [  0/625]  eta: 3:43:29  lr: 0.001447  min_lr: 0.001447  loss: 2.9341 (2.9341)  class_acc: 0.5508 (0.5508)  weight_decay: 0.0500 (0.0500)  time: 21.4560  data: 21.3298  max mem: 2905
Epoch: [185]  [200/625]  eta: 0:14:25  lr: 0.001440  min_lr: 0.001440  loss: 2.9338 (2.9171)  class_acc: 0.5430 (0.5453)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4908 (2.6441)  time: 1.9272  data: 0.0011  max mem: 2905
Epoch: [185]  [400/625]  eta: 0:07:21  lr: 0.001433  min_lr: 0.001433  loss: 2.8433 (2.9245)  class_acc: 0.5430 (0.5437)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4949 (2.6339)  time: 1.9658  data: 0.6593  max mem: 2905
Epoch: [185]  [600/625]  eta: 0:00:48  lr: 0.001426  min_lr: 0.001426  loss: 2.9800 (2.9352)  class_acc: 0.5352 (0.5412)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4372 (2.6900)  time: 1.9224  data: 0.0364  max mem: 2905
Epoch: [185]  [624/625]  eta: 0:00:01  lr: 0.001426  min_lr: 0.001426  loss: 2.9000 (2.9343)  class_acc: 0.5469 (0.5417)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.7563 (2.7122)  time: 0.7465  data: 0.0024  max mem: 2905
Epoch: [185] Total time: 0:19:51 (1.9071 s / it)
Averaged stats: lr: 0.001426  min_lr: 0.001426  loss: 2.9000 (2.9410)  class_acc: 0.5469 (0.5393)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.7563 (2.7122)
Test:  [ 0/50]  eta: 0:11:00  loss: 3.0293 (3.0293)  acc1: 33.6000 (33.6000)  acc5: 63.2000 (63.2000)  time: 13.2031  data: 13.1694  max mem: 2905
Test:  [10/50]  eta: 0:01:35  loss: 2.5866 (2.6063)  acc1: 46.4000 (45.4545)  acc5: 72.0000 (69.3818)  time: 2.3922  data: 2.3695  max mem: 2905
Test:  [20/50]  eta: 0:00:56  loss: 2.6123 (2.7243)  acc1: 44.8000 (43.5048)  acc5: 70.4000 (68.4191)  time: 1.3224  data: 1.3021  max mem: 2905
Test:  [30/50]  eta: 0:00:31  loss: 2.7870 (2.7195)  acc1: 42.4000 (43.7677)  acc5: 68.0000 (68.8000)  time: 1.1298  data: 1.1117  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 2.7870 (2.7756)  acc1: 42.4000 (43.3561)  acc5: 67.2000 (67.8829)  time: 0.6770  data: 0.6586  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.9163 (2.7990)  acc1: 42.4000 (43.2480)  acc5: 65.6000 (67.5040)  time: 0.6251  data: 0.6053  max mem: 2905
Test: Total time: 0:00:53 (1.0721 s / it)
* Acc@1 44.076 Acc@5 68.814 loss 2.750
Accuracy of the model on the 50000 test images: 44.1%
Max accuracy: 50.54%
Epoch: [186]  [  0/625]  eta: 4:09:10  lr: 0.001425  min_lr: 0.001425  loss: 2.8937 (2.8937)  class_acc: 0.5703 (0.5703)  weight_decay: 0.0500 (0.0500)  time: 23.9201  data: 19.2024  max mem: 2905
Epoch: [186]  [200/625]  eta: 0:14:16  lr: 0.001419  min_lr: 0.001419  loss: 2.9270 (2.9184)  class_acc: 0.5352 (0.5439)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7794 (2.8982)  time: 1.9494  data: 0.1967  max mem: 2905
Epoch: [186]  [400/625]  eta: 0:07:22  lr: 0.001412  min_lr: 0.001412  loss: 2.9199 (2.9339)  class_acc: 0.5430 (0.5404)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5156 (2.8205)  time: 2.0454  data: 0.0007  max mem: 2905
Epoch: [186]  [600/625]  eta: 0:00:48  lr: 0.001405  min_lr: 0.001405  loss: 2.9077 (2.9397)  class_acc: 0.5469 (0.5392)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.7382 (2.7667)  time: 1.8180  data: 0.0358  max mem: 2905
Epoch: [186]  [624/625]  eta: 0:00:01  lr: 0.001404  min_lr: 0.001404  loss: 2.9374 (2.9397)  class_acc: 0.5391 (0.5391)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5795 (2.8222)  time: 0.7186  data: 0.0361  max mem: 2905
Epoch: [186] Total time: 0:19:44 (1.8949 s / it)
Averaged stats: lr: 0.001404  min_lr: 0.001404  loss: 2.9374 (2.9394)  class_acc: 0.5391 (0.5399)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5795 (2.8222)
Test:  [ 0/50]  eta: 0:09:55  loss: 3.7056 (3.7056)  acc1: 29.6000 (29.6000)  acc5: 53.6000 (53.6000)  time: 11.9145  data: 11.8874  max mem: 2905
Test:  [10/50]  eta: 0:01:20  loss: 3.1112 (3.0873)  acc1: 38.4000 (40.2182)  acc5: 63.2000 (62.2545)  time: 2.0080  data: 1.9893  max mem: 2905
Test:  [20/50]  eta: 0:00:47  loss: 3.2475 (3.3704)  acc1: 36.8000 (35.0095)  acc5: 61.6000 (59.3905)  time: 1.0534  data: 1.0354  max mem: 2905
Test:  [30/50]  eta: 0:00:27  loss: 3.4997 (3.3361)  acc1: 32.0000 (35.7677)  acc5: 58.4000 (59.8710)  time: 1.0561  data: 1.0359  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 3.3638 (3.3853)  acc1: 35.2000 (35.5707)  acc5: 60.0000 (59.1610)  time: 0.7788  data: 0.7577  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 3.4239 (3.4182)  acc1: 35.2000 (35.4720)  acc5: 55.2000 (58.4640)  time: 0.6766  data: 0.6555  max mem: 2905
Test: Total time: 0:00:49 (0.9874 s / it)
* Acc@1 35.878 Acc@5 58.962 loss 3.370
Accuracy of the model on the 50000 test images: 35.9%
Max accuracy: 50.54%
Epoch: [187]  [  0/625]  eta: 3:44:01  lr: 0.001404  min_lr: 0.001404  loss: 2.9614 (2.9614)  class_acc: 0.5000 (0.5000)  weight_decay: 0.0500 (0.0500)  time: 21.5067  data: 21.3693  max mem: 2905
Epoch: [187]  [200/625]  eta: 0:14:24  lr: 0.001397  min_lr: 0.001397  loss: 2.9544 (2.9256)  class_acc: 0.5391 (0.5437)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.9375 (3.1494)  time: 1.7960  data: 0.0011  max mem: 2905
Epoch: [187]  [400/625]  eta: 0:07:23  lr: 0.001390  min_lr: 0.001390  loss: 2.9764 (2.9318)  class_acc: 0.5312 (0.5414)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9406 (2.9901)  time: 1.8576  data: 0.0146  max mem: 2905
Epoch: [187]  [600/625]  eta: 0:00:49  lr: 0.001383  min_lr: 0.001383  loss: 2.9485 (2.9383)  class_acc: 0.5430 (0.5402)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0422 (inf)  time: 1.9623  data: 0.0188  max mem: 2905
Epoch: [187]  [624/625]  eta: 0:00:01  lr: 0.001383  min_lr: 0.001383  loss: 2.9454 (2.9393)  class_acc: 0.5391 (0.5401)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7637 (inf)  time: 0.7093  data: 0.0014  max mem: 2905
Epoch: [187] Total time: 0:19:56 (1.9146 s / it)
Averaged stats: lr: 0.001383  min_lr: 0.001383  loss: 2.9454 (2.9338)  class_acc: 0.5391 (0.5408)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7637 (inf)
Test:  [ 0/50]  eta: 0:10:12  loss: 3.0444 (3.0444)  acc1: 36.0000 (36.0000)  acc5: 65.6000 (65.6000)  time: 12.2456  data: 12.2174  max mem: 2905
Test:  [10/50]  eta: 0:01:21  loss: 2.8346 (2.8365)  acc1: 41.6000 (41.4545)  acc5: 68.0000 (66.8364)  time: 2.0431  data: 2.0224  max mem: 2905
Test:  [20/50]  eta: 0:00:48  loss: 2.8422 (2.8652)  acc1: 40.0000 (41.3714)  acc5: 68.0000 (66.6286)  time: 1.0931  data: 1.0739  max mem: 2905
Test:  [30/50]  eta: 0:00:28  loss: 2.8550 (2.8851)  acc1: 40.8000 (41.3419)  acc5: 65.6000 (66.0129)  time: 1.0729  data: 1.0530  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 2.9609 (2.9396)  acc1: 40.8000 (40.2927)  acc5: 64.0000 (65.6585)  time: 0.6442  data: 0.6217  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 3.0720 (2.9490)  acc1: 40.0000 (40.1120)  acc5: 64.0000 (65.7280)  time: 0.5637  data: 0.5424  max mem: 2905
Test: Total time: 0:00:47 (0.9553 s / it)
* Acc@1 40.528 Acc@5 65.984 loss 2.923
Accuracy of the model on the 50000 test images: 40.5%
Max accuracy: 50.54%
Epoch: [188]  [  0/625]  eta: 3:45:55  lr: 0.001383  min_lr: 0.001383  loss: 2.9962 (2.9962)  class_acc: 0.5156 (0.5156)  weight_decay: 0.0500 (0.0500)  time: 21.6886  data: 18.3213  max mem: 2905
Epoch: [188]  [200/625]  eta: 0:13:36  lr: 0.001376  min_lr: 0.001376  loss: 2.9010 (2.9230)  class_acc: 0.5469 (0.5462)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.8185 (2.5813)  time: 1.7251  data: 0.0006  max mem: 2905
Epoch: [188]  [400/625]  eta: 0:07:04  lr: 0.001369  min_lr: 0.001369  loss: 2.8965 (2.9250)  class_acc: 0.5469 (0.5444)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.8789 (2.7608)  time: 1.8175  data: 0.0006  max mem: 2905
Epoch: [188]  [600/625]  eta: 0:00:47  lr: 0.001362  min_lr: 0.001362  loss: 2.9332 (2.9299)  class_acc: 0.5312 (0.5440)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.4110 (2.8955)  time: 2.0495  data: 0.0008  max mem: 2905
Epoch: [188]  [624/625]  eta: 0:00:01  lr: 0.001361  min_lr: 0.001361  loss: 2.9181 (2.9307)  class_acc: 0.5352 (0.5440)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.0252 (2.8989)  time: 0.6879  data: 0.0015  max mem: 2905
Epoch: [188] Total time: 0:19:35 (1.8806 s / it)
Averaged stats: lr: 0.001361  min_lr: 0.001361  loss: 2.9181 (2.9294)  class_acc: 0.5352 (0.5421)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.0252 (2.8989)
Test:  [ 0/50]  eta: 0:10:18  loss: 2.4576 (2.4576)  acc1: 47.2000 (47.2000)  acc5: 76.0000 (76.0000)  time: 12.3720  data: 12.3444  max mem: 2905
Test:  [10/50]  eta: 0:01:21  loss: 2.4576 (2.5802)  acc1: 47.2000 (45.9636)  acc5: 72.0000 (70.1091)  time: 2.0392  data: 2.0203  max mem: 2905
Test:  [20/50]  eta: 0:00:47  loss: 2.7938 (2.7306)  acc1: 41.6000 (43.1238)  acc5: 68.0000 (68.3048)  time: 1.0518  data: 1.0338  max mem: 2905
Test:  [30/50]  eta: 0:00:26  loss: 2.7938 (2.7247)  acc1: 43.2000 (43.8194)  acc5: 68.0000 (68.3097)  time: 0.9475  data: 0.9288  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 2.7286 (2.7323)  acc1: 44.0000 (43.4732)  acc5: 68.8000 (68.1171)  time: 0.5953  data: 0.5743  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 2.7033 (2.7306)  acc1: 43.2000 (43.5360)  acc5: 68.8000 (68.0640)  time: 0.5090  data: 0.4883  max mem: 2905
Test: Total time: 0:00:46 (0.9286 s / it)
* Acc@1 44.302 Acc@5 68.832 loss 2.695
Accuracy of the model on the 50000 test images: 44.3%
Max accuracy: 50.54%
Epoch: [189]  [  0/625]  eta: 3:21:22  lr: 0.001361  min_lr: 0.001361  loss: 2.8026 (2.8026)  class_acc: 0.5664 (0.5664)  weight_decay: 0.0500 (0.0500)  time: 19.3315  data: 18.9044  max mem: 2905
Epoch: [189]  [200/625]  eta: 0:14:03  lr: 0.001355  min_lr: 0.001355  loss: 2.8865 (2.9203)  class_acc: 0.5391 (0.5457)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3264 (2.8017)  time: 1.8587  data: 1.5231  max mem: 2905
Epoch: [189]  [400/625]  eta: 0:07:12  lr: 0.001348  min_lr: 0.001348  loss: 2.9280 (2.9199)  class_acc: 0.5352 (0.5448)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2555 (2.7308)  time: 1.8627  data: 1.6581  max mem: 2905
Epoch: [189]  [600/625]  eta: 0:00:47  lr: 0.001341  min_lr: 0.001341  loss: 2.9145 (2.9225)  class_acc: 0.5430 (0.5443)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2839 (2.7543)  time: 1.8939  data: 1.7072  max mem: 2905
Epoch: [189]  [624/625]  eta: 0:00:01  lr: 0.001340  min_lr: 0.001340  loss: 2.8932 (2.9218)  class_acc: 0.5469 (0.5444)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6119 (2.7563)  time: 0.7162  data: 0.5508  max mem: 2905
Epoch: [189] Total time: 0:19:26 (1.8669 s / it)
Averaged stats: lr: 0.001340  min_lr: 0.001340  loss: 2.8932 (2.9291)  class_acc: 0.5469 (0.5422)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6119 (2.7563)
Test:  [ 0/50]  eta: 0:09:56  loss: 2.4711 (2.4711)  acc1: 40.8000 (40.8000)  acc5: 74.4000 (74.4000)  time: 11.9284  data: 11.9016  max mem: 2905
Test:  [10/50]  eta: 0:01:21  loss: 2.3305 (2.2734)  acc1: 51.2000 (50.7636)  acc5: 74.4000 (74.1818)  time: 2.0462  data: 2.0261  max mem: 2905
Test:  [20/50]  eta: 0:00:48  loss: 2.3902 (2.3994)  acc1: 48.8000 (48.4571)  acc5: 73.6000 (74.0191)  time: 1.1108  data: 1.0915  max mem: 2905
Test:  [30/50]  eta: 0:00:28  loss: 2.4373 (2.4211)  acc1: 44.8000 (48.2065)  acc5: 73.6000 (73.5742)  time: 1.0868  data: 1.0677  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 2.4373 (2.4279)  acc1: 48.8000 (48.8390)  acc5: 74.4000 (73.7951)  time: 0.6950  data: 0.6742  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 2.3989 (2.4369)  acc1: 48.8000 (48.7520)  acc5: 74.4000 (73.7280)  time: 0.5423  data: 0.5211  max mem: 2905
Test: Total time: 0:00:49 (0.9888 s / it)
* Acc@1 49.300 Acc@5 74.036 loss 2.405
Accuracy of the model on the 50000 test images: 49.3%
Max accuracy: 50.54%
Epoch: [190]  [  0/625]  eta: 3:32:30  lr: 0.001340  min_lr: 0.001340  loss: 2.9151 (2.9151)  class_acc: 0.5547 (0.5547)  weight_decay: 0.0500 (0.0500)  time: 20.4012  data: 20.0372  max mem: 2905
Epoch: [190]  [200/625]  eta: 0:14:04  lr: 0.001333  min_lr: 0.001333  loss: 2.9531 (2.9229)  class_acc: 0.5391 (0.5451)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5471 (2.8818)  time: 2.0116  data: 0.1104  max mem: 2905
Epoch: [190]  [400/625]  eta: 0:07:22  lr: 0.001327  min_lr: 0.001327  loss: 2.9694 (2.9240)  class_acc: 0.5430 (0.5434)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.4676 (3.0408)  time: 1.9912  data: 0.0081  max mem: 2905
Epoch: [190]  [600/625]  eta: 0:00:49  lr: 0.001320  min_lr: 0.001320  loss: 2.9507 (2.9288)  class_acc: 0.5352 (0.5424)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2835 (2.9434)  time: 2.0388  data: 0.0008  max mem: 2905
Epoch: [190]  [624/625]  eta: 0:00:01  lr: 0.001319  min_lr: 0.001319  loss: 2.9223 (2.9306)  class_acc: 0.5352 (0.5420)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.8739 (2.9507)  time: 0.8172  data: 0.0081  max mem: 2905
Epoch: [190] Total time: 0:20:08 (1.9339 s / it)
Averaged stats: lr: 0.001319  min_lr: 0.001319  loss: 2.9223 (2.9253)  class_acc: 0.5352 (0.5428)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.8739 (2.9507)
Test:  [ 0/50]  eta: 0:10:18  loss: 2.5478 (2.5478)  acc1: 45.6000 (45.6000)  acc5: 72.0000 (72.0000)  time: 12.3603  data: 12.3302  max mem: 2905
Test:  [10/50]  eta: 0:01:17  loss: 2.4530 (2.4388)  acc1: 49.6000 (49.5273)  acc5: 72.0000 (71.7818)  time: 1.9423  data: 1.9225  max mem: 2905
Test:  [20/50]  eta: 0:00:44  loss: 2.5764 (2.6544)  acc1: 46.4000 (45.2571)  acc5: 69.6000 (70.0191)  time: 0.9527  data: 0.9328  max mem: 2905
Test:  [30/50]  eta: 0:00:26  loss: 2.7053 (2.6843)  acc1: 40.8000 (45.0323)  acc5: 69.6000 (69.7032)  time: 0.9586  data: 0.9387  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 2.5968 (2.6682)  acc1: 44.0000 (44.9951)  acc5: 71.2000 (70.0098)  time: 0.9298  data: 0.9107  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.5268 (2.6248)  acc1: 45.6000 (45.6960)  acc5: 72.8000 (70.8320)  time: 0.8134  data: 0.7944  max mem: 2905
Test: Total time: 0:00:54 (1.0879 s / it)
* Acc@1 46.152 Acc@5 70.916 loss 2.604
Accuracy of the model on the 50000 test images: 46.2%
Max accuracy: 50.54%
Epoch: [191]  [  0/625]  eta: 3:45:55  lr: 0.001319  min_lr: 0.001319  loss: 2.9978 (2.9978)  class_acc: 0.5508 (0.5508)  weight_decay: 0.0500 (0.0500)  time: 21.6880  data: 20.7601  max mem: 2905
Epoch: [191]  [200/625]  eta: 0:13:59  lr: 0.001312  min_lr: 0.001312  loss: 2.9575 (2.9073)  class_acc: 0.5273 (0.5475)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.0668 (2.9456)  time: 1.9654  data: 0.1931  max mem: 2905
Epoch: [191]  [400/625]  eta: 0:07:12  lr: 0.001305  min_lr: 0.001305  loss: 2.9726 (2.9155)  class_acc: 0.5391 (0.5456)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.9952 (3.0188)  time: 1.8039  data: 0.0006  max mem: 2905
Epoch: [191]  [600/625]  eta: 0:00:48  lr: 0.001299  min_lr: 0.001299  loss: 2.9182 (2.9250)  class_acc: 0.5391 (0.5435)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.7701 (3.0389)  time: 1.8028  data: 0.0008  max mem: 2905
Epoch: [191]  [624/625]  eta: 0:00:01  lr: 0.001298  min_lr: 0.001298  loss: 2.9384 (2.9255)  class_acc: 0.5312 (0.5433)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.3549 (3.0748)  time: 0.6599  data: 0.0056  max mem: 2905
Epoch: [191] Total time: 0:19:49 (1.9032 s / it)
Averaged stats: lr: 0.001298  min_lr: 0.001298  loss: 2.9384 (2.9227)  class_acc: 0.5312 (0.5436)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.3549 (3.0748)
Test:  [ 0/50]  eta: 0:09:42  loss: 2.9738 (2.9738)  acc1: 44.0000 (44.0000)  acc5: 65.6000 (65.6000)  time: 11.6522  data: 11.6238  max mem: 2905
Test:  [10/50]  eta: 0:01:16  loss: 2.6974 (2.6458)  acc1: 45.6000 (46.8364)  acc5: 71.2000 (71.0545)  time: 1.9167  data: 1.8939  max mem: 2905
Test:  [20/50]  eta: 0:00:42  loss: 2.7606 (2.8378)  acc1: 41.6000 (43.0476)  acc5: 68.8000 (67.6571)  time: 0.8942  data: 0.8732  max mem: 2905
Test:  [30/50]  eta: 0:00:24  loss: 2.8132 (2.8180)  acc1: 40.0000 (43.1484)  acc5: 64.8000 (67.9742)  time: 0.8142  data: 0.7939  max mem: 2905
Test:  [40/50]  eta: 0:00:10  loss: 2.7896 (2.8379)  acc1: 42.4000 (42.8293)  acc5: 68.0000 (67.7268)  time: 0.6545  data: 0.6347  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 2.7911 (2.8367)  acc1: 41.6000 (42.7040)  acc5: 67.2000 (67.6960)  time: 0.5751  data: 0.5550  max mem: 2905
Test: Total time: 0:00:45 (0.9002 s / it)
* Acc@1 43.686 Acc@5 67.946 loss 2.803
Accuracy of the model on the 50000 test images: 43.7%
Max accuracy: 50.54%
Epoch: [192]  [  0/625]  eta: 3:29:30  lr: 0.001298  min_lr: 0.001298  loss: 2.7630 (2.7630)  class_acc: 0.5977 (0.5977)  weight_decay: 0.0500 (0.0500)  time: 20.1135  data: 18.1851  max mem: 2905
Epoch: [192]  [200/625]  eta: 0:14:01  lr: 0.001291  min_lr: 0.001291  loss: 2.9235 (2.9231)  class_acc: 0.5430 (0.5437)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.9417 (2.8943)  time: 1.9627  data: 0.0010  max mem: 2905
Epoch: [192]  [400/625]  eta: 0:07:28  lr: 0.001284  min_lr: 0.001284  loss: 2.9102 (2.9212)  class_acc: 0.5430 (0.5442)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0723 (2.8096)  time: 2.0547  data: 0.0008  max mem: 2905
Epoch: [192]  [600/625]  eta: 0:00:49  lr: 0.001278  min_lr: 0.001278  loss: 2.9390 (2.9262)  class_acc: 0.5312 (0.5434)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0961 (2.7313)  time: 1.9570  data: 0.0008  max mem: 2905
Epoch: [192]  [624/625]  eta: 0:00:01  lr: 0.001277  min_lr: 0.001277  loss: 2.8979 (2.9264)  class_acc: 0.5391 (0.5431)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3891 (2.7369)  time: 0.6761  data: 0.0016  max mem: 2905
Epoch: [192] Total time: 0:20:32 (1.9717 s / it)
Averaged stats: lr: 0.001277  min_lr: 0.001277  loss: 2.8979 (2.9222)  class_acc: 0.5391 (0.5438)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3891 (2.7369)
Test:  [ 0/50]  eta: 0:10:33  loss: 3.0853 (3.0853)  acc1: 36.0000 (36.0000)  acc5: 66.4000 (66.4000)  time: 12.6765  data: 12.6396  max mem: 2905
Test:  [10/50]  eta: 0:01:17  loss: 3.2519 (3.1364)  acc1: 36.0000 (39.2000)  acc5: 64.0000 (63.3455)  time: 1.9456  data: 1.9255  max mem: 2905
Test:  [20/50]  eta: 0:00:44  loss: 3.3824 (3.2797)  acc1: 35.2000 (36.8000)  acc5: 59.2000 (61.7524)  time: 0.9165  data: 0.8971  max mem: 2905
Test:  [30/50]  eta: 0:00:27  loss: 3.3651 (3.2682)  acc1: 35.2000 (37.5742)  acc5: 59.2000 (61.5742)  time: 1.0402  data: 1.0209  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 3.1933 (3.2724)  acc1: 35.2000 (36.9366)  acc5: 60.8000 (61.6195)  time: 1.1026  data: 1.0839  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 3.1658 (3.2495)  acc1: 35.2000 (36.7840)  acc5: 60.8000 (61.6320)  time: 0.7949  data: 0.7734  max mem: 2905
Test: Total time: 0:00:58 (1.1661 s / it)
* Acc@1 37.428 Acc@5 62.086 loss 3.206
Accuracy of the model on the 50000 test images: 37.4%
Max accuracy: 50.54%
Epoch: [193]  [  0/625]  eta: 3:25:24  lr: 0.001277  min_lr: 0.001277  loss: 2.8153 (2.8153)  class_acc: 0.5586 (0.5586)  weight_decay: 0.0500 (0.0500)  time: 19.7195  data: 19.0394  max mem: 2905
Epoch: [193]  [200/625]  eta: 0:14:31  lr: 0.001270  min_lr: 0.001270  loss: 2.9840 (2.9127)  class_acc: 0.5430 (0.5435)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.8202 (3.0284)  time: 1.9617  data: 0.0010  max mem: 2905
Epoch: [193]  [400/625]  eta: 0:07:32  lr: 0.001264  min_lr: 0.001264  loss: 2.8956 (2.9230)  class_acc: 0.5508 (0.5431)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1415 (2.7949)  time: 1.9404  data: 0.0012  max mem: 2905
Epoch: [193]  [600/625]  eta: 0:00:49  lr: 0.001257  min_lr: 0.001257  loss: 2.8523 (2.9272)  class_acc: 0.5625 (0.5412)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4802 (2.9189)  time: 2.0040  data: 0.0011  max mem: 2905
Epoch: [193]  [624/625]  eta: 0:00:01  lr: 0.001256  min_lr: 0.001256  loss: 2.9145 (2.9264)  class_acc: 0.5352 (0.5415)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.9187 (2.9353)  time: 0.7411  data: 0.0022  max mem: 2905
Epoch: [193] Total time: 0:20:18 (1.9492 s / it)
Averaged stats: lr: 0.001256  min_lr: 0.001256  loss: 2.9145 (2.9196)  class_acc: 0.5352 (0.5440)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.9187 (2.9353)
Test:  [ 0/50]  eta: 0:10:23  loss: 3.2407 (3.2407)  acc1: 31.2000 (31.2000)  acc5: 63.2000 (63.2000)  time: 12.4640  data: 12.4337  max mem: 2905
Test:  [10/50]  eta: 0:01:18  loss: 2.9510 (2.9637)  acc1: 39.2000 (38.8364)  acc5: 64.8000 (64.8000)  time: 1.9627  data: 1.9430  max mem: 2905
Test:  [20/50]  eta: 0:00:45  loss: 3.0986 (3.1029)  acc1: 37.6000 (36.6857)  acc5: 63.2000 (63.1238)  time: 0.9596  data: 0.9403  max mem: 2905
Test:  [30/50]  eta: 0:00:26  loss: 3.1853 (3.0552)  acc1: 36.8000 (37.7548)  acc5: 61.6000 (63.4065)  time: 0.9639  data: 0.9449  max mem: 2905
Test:  [40/50]  eta: 0:00:10  loss: 3.0145 (3.0531)  acc1: 39.2000 (38.3610)  acc5: 61.6000 (63.6683)  time: 0.6666  data: 0.6473  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 2.9667 (3.0554)  acc1: 39.2000 (38.3680)  acc5: 65.6000 (63.9040)  time: 0.5243  data: 0.5049  max mem: 2905
Test: Total time: 0:00:47 (0.9570 s / it)
* Acc@1 39.556 Acc@5 64.008 loss 3.017
Accuracy of the model on the 50000 test images: 39.6%
Max accuracy: 50.54%
Epoch: [194]  [  0/625]  eta: 3:21:16  lr: 0.001256  min_lr: 0.001256  loss: 2.8731 (2.8731)  class_acc: 0.5391 (0.5391)  weight_decay: 0.0500 (0.0500)  time: 19.3229  data: 19.1959  max mem: 2905
Epoch: [194]  [200/625]  eta: 0:14:15  lr: 0.001249  min_lr: 0.001249  loss: 2.9252 (2.9092)  class_acc: 0.5508 (0.5444)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5505 (inf)  time: 1.9251  data: 0.1120  max mem: 2905
Epoch: [194]  [400/625]  eta: 0:07:20  lr: 0.001243  min_lr: 0.001243  loss: 2.9534 (2.9101)  class_acc: 0.5312 (0.5454)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3721 (inf)  time: 2.0637  data: 0.0744  max mem: 2905
Epoch: [194]  [600/625]  eta: 0:00:49  lr: 0.001236  min_lr: 0.001236  loss: 2.8972 (2.9143)  class_acc: 0.5391 (0.5447)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9056 (inf)  time: 2.0990  data: 0.0424  max mem: 2905
Epoch: [194]  [624/625]  eta: 0:00:01  lr: 0.001235  min_lr: 0.001235  loss: 2.9051 (2.9143)  class_acc: 0.5352 (0.5445)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0697 (inf)  time: 0.7116  data: 0.0019  max mem: 2905
Epoch: [194] Total time: 0:19:57 (1.9156 s / it)
Averaged stats: lr: 0.001235  min_lr: 0.001235  loss: 2.9051 (2.9127)  class_acc: 0.5352 (0.5459)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0697 (inf)
Test:  [ 0/50]  eta: 0:10:12  loss: 2.5322 (2.5322)  acc1: 49.6000 (49.6000)  acc5: 72.8000 (72.8000)  time: 12.2496  data: 12.2186  max mem: 2905
Test:  [10/50]  eta: 0:01:13  loss: 2.4090 (2.4054)  acc1: 49.6000 (50.0364)  acc5: 73.6000 (73.0182)  time: 1.8260  data: 1.8060  max mem: 2905
Test:  [20/50]  eta: 0:00:40  loss: 2.5546 (2.5443)  acc1: 46.4000 (46.9333)  acc5: 70.4000 (71.0857)  time: 0.8114  data: 0.7925  max mem: 2905
Test:  [30/50]  eta: 0:00:24  loss: 2.5546 (2.5469)  acc1: 44.8000 (46.8903)  acc5: 69.6000 (71.3290)  time: 0.9007  data: 0.8816  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 2.4524 (2.5893)  acc1: 47.2000 (46.6146)  acc5: 69.6000 (70.5756)  time: 0.8979  data: 0.8789  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 2.4173 (2.5784)  acc1: 47.2000 (46.7840)  acc5: 71.2000 (70.8480)  time: 0.5608  data: 0.5419  max mem: 2905
Test: Total time: 0:00:49 (0.9893 s / it)
* Acc@1 46.720 Acc@5 71.438 loss 2.548
Accuracy of the model on the 50000 test images: 46.7%
Max accuracy: 50.54%
Epoch: [195]  [  0/625]  eta: 3:50:55  lr: 0.001235  min_lr: 0.001235  loss: 2.9635 (2.9635)  class_acc: 0.5039 (0.5039)  weight_decay: 0.0500 (0.0500)  time: 22.1691  data: 17.6797  max mem: 2905
Epoch: [195]  [200/625]  eta: 0:14:16  lr: 0.001229  min_lr: 0.001229  loss: 2.9411 (2.8931)  class_acc: 0.5352 (0.5508)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.1960 (3.4029)  time: 1.9323  data: 1.6963  max mem: 2905
Epoch: [195]  [400/625]  eta: 0:07:21  lr: 0.001222  min_lr: 0.001222  loss: 2.8517 (2.9087)  class_acc: 0.5430 (0.5473)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4624 (3.3010)  time: 1.9755  data: 1.1315  max mem: 2905
Epoch: [195]  [600/625]  eta: 0:00:51  lr: 0.001215  min_lr: 0.001215  loss: 2.9085 (2.9107)  class_acc: 0.5391 (0.5471)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9755 (3.0550)  time: 2.3877  data: 0.0006  max mem: 2905
Epoch: [195]  [624/625]  eta: 0:00:02  lr: 0.001215  min_lr: 0.001215  loss: 2.8800 (2.9105)  class_acc: 0.5430 (0.5469)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7811 (3.0690)  time: 1.3988  data: 0.0079  max mem: 2905
Epoch: [195] Total time: 0:21:17 (2.0439 s / it)
Averaged stats: lr: 0.001215  min_lr: 0.001215  loss: 2.8800 (2.9122)  class_acc: 0.5430 (0.5458)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7811 (3.0690)
Test:  [ 0/50]  eta: 0:25:24  loss: 2.5202 (2.5202)  acc1: 42.4000 (42.4000)  acc5: 73.6000 (73.6000)  time: 30.4993  data: 30.4583  max mem: 2905
Test:  [10/50]  eta: 0:02:49  loss: 2.5202 (2.4166)  acc1: 47.2000 (48.2909)  acc5: 73.6000 (73.9636)  time: 4.2459  data: 4.2237  max mem: 2905
Test:  [20/50]  eta: 0:01:26  loss: 2.5572 (2.5614)  acc1: 45.6000 (45.2190)  acc5: 73.6000 (71.8476)  time: 1.5077  data: 1.4876  max mem: 2905
Test:  [30/50]  eta: 0:00:47  loss: 2.6460 (2.5833)  acc1: 45.6000 (45.5484)  acc5: 68.8000 (71.2258)  time: 1.3445  data: 1.3242  max mem: 2905
Test:  [40/50]  eta: 0:00:19  loss: 2.6460 (2.5963)  acc1: 45.6000 (45.7171)  acc5: 68.8000 (70.9073)  time: 1.0184  data: 0.9981  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.6124 (2.6062)  acc1: 45.6000 (45.7760)  acc5: 68.8000 (70.5920)  time: 0.8597  data: 0.8377  max mem: 2905
Test: Total time: 0:01:22 (1.6425 s / it)
* Acc@1 46.208 Acc@5 71.248 loss 2.572
Accuracy of the model on the 50000 test images: 46.2%
Max accuracy: 50.54%
Epoch: [196]  [  0/625]  eta: 3:31:25  lr: 0.001215  min_lr: 0.001215  loss: 2.7980 (2.7980)  class_acc: 0.5938 (0.5938)  weight_decay: 0.0500 (0.0500)  time: 20.2969  data: 17.0873  max mem: 2905
Epoch: [196]  [200/625]  eta: 0:14:25  lr: 0.001208  min_lr: 0.001208  loss: 2.9126 (2.9048)  class_acc: 0.5586 (0.5485)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0398 (3.0081)  time: 1.7068  data: 0.0381  max mem: 2905
Epoch: [196]  [400/625]  eta: 0:07:22  lr: 0.001201  min_lr: 0.001201  loss: 2.9202 (2.9100)  class_acc: 0.5430 (0.5459)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.8460 (2.9066)  time: 1.8193  data: 0.0122  max mem: 2905
Epoch: [196]  [600/625]  eta: 0:00:48  lr: 0.001195  min_lr: 0.001195  loss: 2.9258 (2.9145)  class_acc: 0.5391 (0.5452)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.9828 (inf)  time: 1.9735  data: 0.0007  max mem: 2905
Epoch: [196]  [624/625]  eta: 0:00:01  lr: 0.001194  min_lr: 0.001194  loss: 2.9119 (2.9149)  class_acc: 0.5391 (0.5452)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0852 (inf)  time: 0.6276  data: 0.0013  max mem: 2905
Epoch: [196] Total time: 0:19:53 (1.9094 s / it)
Averaged stats: lr: 0.001194  min_lr: 0.001194  loss: 2.9119 (2.9102)  class_acc: 0.5391 (0.5460)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0852 (inf)
Test:  [ 0/50]  eta: 0:09:40  loss: 2.9230 (2.9230)  acc1: 43.2000 (43.2000)  acc5: 64.0000 (64.0000)  time: 11.6038  data: 11.5685  max mem: 2905
Test:  [10/50]  eta: 0:01:17  loss: 2.7042 (2.7395)  acc1: 44.0000 (42.4727)  acc5: 70.4000 (68.7273)  time: 1.9352  data: 1.9158  max mem: 2905
Test:  [20/50]  eta: 0:00:45  loss: 2.8235 (2.8646)  acc1: 42.4000 (41.6381)  acc5: 68.8000 (67.4667)  time: 1.0150  data: 0.9971  max mem: 2905
Test:  [30/50]  eta: 0:00:26  loss: 2.9258 (2.8937)  acc1: 40.0000 (41.6000)  acc5: 64.0000 (66.7871)  time: 0.9958  data: 0.9764  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 2.9333 (2.9197)  acc1: 40.0000 (41.5024)  acc5: 64.0000 (66.2244)  time: 0.7575  data: 0.7353  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 2.9671 (2.9522)  acc1: 37.6000 (40.9600)  acc5: 64.0000 (65.6320)  time: 0.6845  data: 0.6635  max mem: 2905
Test: Total time: 0:00:48 (0.9646 s / it)
* Acc@1 42.232 Acc@5 65.968 loss 2.912
Accuracy of the model on the 50000 test images: 42.2%
Max accuracy: 50.54%
Epoch: [197]  [  0/625]  eta: 3:29:50  lr: 0.001194  min_lr: 0.001194  loss: 2.9084 (2.9084)  class_acc: 0.5742 (0.5742)  weight_decay: 0.0500 (0.0500)  time: 20.1449  data: 18.3367  max mem: 2905
Epoch: [197]  [200/625]  eta: 0:14:14  lr: 0.001187  min_lr: 0.001187  loss: 2.9193 (2.8950)  class_acc: 0.5391 (0.5513)  weight_decay: 0.0500 (0.0500)  grad_norm: 4.0277 (3.1640)  time: 2.3592  data: 0.0009  max mem: 2905
Epoch: [197]  [400/625]  eta: 0:07:22  lr: 0.001181  min_lr: 0.001181  loss: 2.9294 (2.9067)  class_acc: 0.5312 (0.5476)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.6728 (3.0156)  time: 2.0081  data: 0.0439  max mem: 2905
Epoch: [197]  [600/625]  eta: 0:00:48  lr: 0.001174  min_lr: 0.001174  loss: 2.8982 (2.9108)  class_acc: 0.5508 (0.5471)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.9043 (2.9878)  time: 1.9745  data: 0.0012  max mem: 2905
Epoch: [197]  [624/625]  eta: 0:00:01  lr: 0.001174  min_lr: 0.001174  loss: 2.8929 (2.9104)  class_acc: 0.5469 (0.5472)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.5255 (3.0085)  time: 0.8097  data: 0.0018  max mem: 2905
Epoch: [197] Total time: 0:19:52 (1.9079 s / it)
Averaged stats: lr: 0.001174  min_lr: 0.001174  loss: 2.8929 (2.9062)  class_acc: 0.5469 (0.5476)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.5255 (3.0085)
Test:  [ 0/50]  eta: 0:10:39  loss: 2.6053 (2.6053)  acc1: 48.0000 (48.0000)  acc5: 71.2000 (71.2000)  time: 12.7980  data: 12.7709  max mem: 2905
Test:  [10/50]  eta: 0:01:25  loss: 2.6053 (2.7062)  acc1: 45.6000 (45.3818)  acc5: 68.0000 (68.5818)  time: 2.1443  data: 2.1248  max mem: 2905
Test:  [20/50]  eta: 0:00:49  loss: 2.7418 (2.7252)  acc1: 42.4000 (43.6952)  acc5: 68.0000 (69.1429)  time: 1.0976  data: 1.0793  max mem: 2905
Test:  [30/50]  eta: 0:00:29  loss: 2.7651 (2.7647)  acc1: 42.4000 (43.6645)  acc5: 68.0000 (68.2839)  time: 1.0932  data: 1.0741  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 2.8325 (2.7633)  acc1: 42.4000 (43.5317)  acc5: 65.6000 (68.2927)  time: 0.8551  data: 0.8357  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.7144 (2.7570)  acc1: 44.8000 (43.8880)  acc5: 68.0000 (68.1920)  time: 0.7523  data: 0.7331  max mem: 2905
Test: Total time: 0:00:55 (1.1199 s / it)
* Acc@1 44.326 Acc@5 68.906 loss 2.720
Accuracy of the model on the 50000 test images: 44.3%
Max accuracy: 50.54%
Epoch: [198]  [  0/625]  eta: 3:44:17  lr: 0.001174  min_lr: 0.001174  loss: 2.8976 (2.8976)  class_acc: 0.5586 (0.5586)  weight_decay: 0.0500 (0.0500)  time: 21.5314  data: 21.1077  max mem: 2905
Epoch: [198]  [200/625]  eta: 0:14:15  lr: 0.001167  min_lr: 0.001167  loss: 2.9582 (2.8968)  class_acc: 0.5352 (0.5524)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.1012 (2.8716)  time: 1.7730  data: 0.0092  max mem: 2905
Epoch: [198]  [400/625]  eta: 0:07:18  lr: 0.001161  min_lr: 0.001161  loss: 2.9451 (2.9057)  class_acc: 0.5352 (0.5483)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2368 (2.9838)  time: 1.8771  data: 0.0006  max mem: 2905
Epoch: [198]  [600/625]  eta: 0:00:48  lr: 0.001154  min_lr: 0.001154  loss: 2.9146 (2.9116)  class_acc: 0.5508 (0.5468)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6784 (2.9849)  time: 1.9182  data: 0.0316  max mem: 2905
Epoch: [198]  [624/625]  eta: 0:00:01  lr: 0.001153  min_lr: 0.001153  loss: 2.8847 (2.9107)  class_acc: 0.5586 (0.5471)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2280 (2.9756)  time: 0.5794  data: 0.0016  max mem: 2905
Epoch: [198] Total time: 0:20:02 (1.9242 s / it)
Averaged stats: lr: 0.001153  min_lr: 0.001153  loss: 2.8847 (2.9039)  class_acc: 0.5586 (0.5482)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2280 (2.9756)
Test:  [ 0/50]  eta: 0:09:40  loss: 2.0337 (2.0337)  acc1: 56.0000 (56.0000)  acc5: 83.2000 (83.2000)  time: 11.6086  data: 11.5774  max mem: 2905
Test:  [10/50]  eta: 0:01:19  loss: 2.1642 (2.2359)  acc1: 50.4000 (51.4182)  acc5: 76.8000 (77.1636)  time: 1.9931  data: 1.9729  max mem: 2905
Test:  [20/50]  eta: 0:00:48  loss: 2.4039 (2.3975)  acc1: 46.4000 (48.2667)  acc5: 75.2000 (74.5905)  time: 1.1121  data: 1.0927  max mem: 2905
Test:  [30/50]  eta: 0:00:28  loss: 2.5155 (2.3722)  acc1: 45.6000 (48.5161)  acc5: 72.8000 (74.5548)  time: 1.1403  data: 1.1219  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 2.3333 (2.3794)  acc1: 48.0000 (48.7610)  acc5: 74.4000 (74.4585)  time: 0.8752  data: 0.8564  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.3231 (2.3685)  acc1: 48.0000 (48.6720)  acc5: 75.2000 (74.6720)  time: 0.7886  data: 0.7674  max mem: 2905
Test: Total time: 0:00:52 (1.0486 s / it)
* Acc@1 49.434 Acc@5 74.702 loss 2.352
Accuracy of the model on the 50000 test images: 49.4%
Max accuracy: 50.54%
Epoch: [199]  [  0/625]  eta: 3:24:01  lr: 0.001153  min_lr: 0.001153  loss: 2.8251 (2.8251)  class_acc: 0.5703 (0.5703)  weight_decay: 0.0500 (0.0500)  time: 19.5865  data: 19.1593  max mem: 2905
Epoch: [199]  [200/625]  eta: 0:14:12  lr: 0.001147  min_lr: 0.001147  loss: 2.9107 (2.8996)  class_acc: 0.5508 (0.5489)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1446 (2.8866)  time: 2.1043  data: 0.0006  max mem: 2905
Epoch: [199]  [400/625]  eta: 0:07:18  lr: 0.001140  min_lr: 0.001140  loss: 2.8977 (2.8954)  class_acc: 0.5312 (0.5497)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2082 (2.9519)  time: 1.9804  data: 0.0259  max mem: 2905
Epoch: [199]  [600/625]  eta: 0:00:47  lr: 0.001134  min_lr: 0.001134  loss: 2.8967 (2.9011)  class_acc: 0.5430 (0.5485)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.4360 (3.0290)  time: 1.9518  data: 0.0006  max mem: 2905
Epoch: [199]  [624/625]  eta: 0:00:01  lr: 0.001133  min_lr: 0.001133  loss: 2.9155 (2.9021)  class_acc: 0.5312 (0.5480)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6394 (3.0202)  time: 1.1443  data: 0.0012  max mem: 2905
Epoch: [199] Total time: 0:19:38 (1.8855 s / it)
Averaged stats: lr: 0.001133  min_lr: 0.001133  loss: 2.9155 (2.9002)  class_acc: 0.5312 (0.5487)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6394 (3.0202)
Test:  [ 0/50]  eta: 0:09:23  loss: 2.4074 (2.4074)  acc1: 47.2000 (47.2000)  acc5: 72.8000 (72.8000)  time: 11.2634  data: 11.2332  max mem: 2905
Test:  [10/50]  eta: 0:01:15  loss: 2.4074 (2.3818)  acc1: 50.4000 (48.3636)  acc5: 72.8000 (72.8727)  time: 1.8962  data: 1.8764  max mem: 2905
Test:  [20/50]  eta: 0:00:43  loss: 2.4303 (2.4626)  acc1: 45.6000 (46.4000)  acc5: 72.8000 (72.9905)  time: 0.9518  data: 0.9322  max mem: 2905
Test:  [30/50]  eta: 0:00:24  loss: 2.5021 (2.5080)  acc1: 44.8000 (46.3226)  acc5: 72.0000 (72.1806)  time: 0.8482  data: 0.8285  max mem: 2905
Test:  [40/50]  eta: 0:00:10  loss: 2.5666 (2.5450)  acc1: 44.8000 (46.0878)  acc5: 68.0000 (71.8439)  time: 0.5616  data: 0.5422  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 2.5415 (2.5402)  acc1: 44.0000 (45.9840)  acc5: 72.8000 (72.0960)  time: 0.5596  data: 0.5396  max mem: 2905
Test: Total time: 0:00:44 (0.8948 s / it)
* Acc@1 47.292 Acc@5 72.470 loss 2.509
Accuracy of the model on the 50000 test images: 47.3%
Max accuracy: 50.54%
Epoch: [200]  [  0/625]  eta: 3:43:06  lr: 0.001133  min_lr: 0.001133  loss: 3.0597 (3.0597)  class_acc: 0.5352 (0.5352)  weight_decay: 0.0500 (0.0500)  time: 21.4189  data: 16.5883  max mem: 2905
Epoch: [200]  [200/625]  eta: 0:13:43  lr: 0.001126  min_lr: 0.001126  loss: 2.8482 (2.8782)  class_acc: 0.5430 (0.5514)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4542 (2.9941)  time: 1.8208  data: 0.0005  max mem: 2905
Epoch: [200]  [400/625]  eta: 0:07:14  lr: 0.001120  min_lr: 0.001120  loss: 2.9281 (2.8937)  class_acc: 0.5469 (0.5493)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7431 (2.9786)  time: 1.9645  data: 0.0006  max mem: 2905
Epoch: [200]  [600/625]  eta: 0:00:48  lr: 0.001114  min_lr: 0.001114  loss: 2.9436 (2.8957)  class_acc: 0.5312 (0.5486)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2007 (3.0581)  time: 1.9204  data: 0.0008  max mem: 2905
Epoch: [200]  [624/625]  eta: 0:00:01  lr: 0.001113  min_lr: 0.001113  loss: 2.9184 (2.8962)  class_acc: 0.5469 (0.5488)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3034 (3.0387)  time: 0.6089  data: 0.0015  max mem: 2905
Epoch: [200] Total time: 0:19:51 (1.9064 s / it)
Averaged stats: lr: 0.001113  min_lr: 0.001113  loss: 2.9184 (2.8964)  class_acc: 0.5469 (0.5493)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3034 (3.0387)
Test:  [ 0/50]  eta: 0:09:40  loss: 2.3553 (2.3553)  acc1: 46.4000 (46.4000)  acc5: 77.6000 (77.6000)  time: 11.6034  data: 11.5752  max mem: 2905
Test:  [10/50]  eta: 0:01:13  loss: 1.9268 (2.0257)  acc1: 55.2000 (56.2182)  acc5: 77.6000 (78.2545)  time: 1.8302  data: 1.8099  max mem: 2905
Test:  [20/50]  eta: 0:00:41  loss: 2.3581 (2.2748)  acc1: 51.2000 (51.3143)  acc5: 74.4000 (75.0476)  time: 0.8820  data: 0.8624  max mem: 2905
Test:  [30/50]  eta: 0:00:25  loss: 2.3483 (2.2302)  acc1: 51.2000 (51.8968)  acc5: 73.6000 (75.8968)  time: 0.9574  data: 0.9384  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 2.1894 (2.2644)  acc1: 52.0000 (51.5122)  acc5: 74.4000 (75.2781)  time: 0.8931  data: 0.8736  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.2296 (2.2747)  acc1: 48.8000 (51.3280)  acc5: 74.4000 (75.3280)  time: 0.5499  data: 0.5294  max mem: 2905
Test: Total time: 0:00:50 (1.0084 s / it)
* Acc@1 52.052 Acc@5 76.100 loss 2.228
Accuracy of the model on the 50000 test images: 52.1%
Max accuracy: 52.05%
Epoch: [201]  [  0/625]  eta: 4:03:00  lr: 0.001113  min_lr: 0.001113  loss: 2.9127 (2.9127)  class_acc: 0.5547 (0.5547)  weight_decay: 0.0500 (0.0500)  time: 23.3292  data: 20.7101  max mem: 2905
Epoch: [201]  [200/625]  eta: 0:14:03  lr: 0.001106  min_lr: 0.001106  loss: 2.9559 (2.8961)  class_acc: 0.5352 (0.5489)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.1042 (3.2214)  time: 2.0185  data: 0.2712  max mem: 2905
Epoch: [201]  [400/625]  eta: 0:07:24  lr: 0.001100  min_lr: 0.001100  loss: 2.8621 (2.9003)  class_acc: 0.5547 (0.5484)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3592 (3.0595)  time: 2.2133  data: 0.0951  max mem: 2905
Epoch: [201]  [600/625]  eta: 0:00:49  lr: 0.001094  min_lr: 0.001094  loss: 2.8697 (2.8973)  class_acc: 0.5547 (0.5489)  weight_decay: 0.0500 (0.0500)  grad_norm: 4.9074 (3.2205)  time: 1.9500  data: 0.0859  max mem: 2905
Epoch: [201]  [624/625]  eta: 0:00:01  lr: 0.001093  min_lr: 0.001093  loss: 2.9323 (2.8977)  class_acc: 0.5547 (0.5490)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.9555 (3.2072)  time: 0.9491  data: 0.0527  max mem: 2905
Epoch: [201] Total time: 0:20:08 (1.9336 s / it)
Averaged stats: lr: 0.001093  min_lr: 0.001093  loss: 2.9323 (2.8937)  class_acc: 0.5547 (0.5502)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.9555 (3.2072)
Test:  [ 0/50]  eta: 0:10:14  loss: 2.6657 (2.6657)  acc1: 46.4000 (46.4000)  acc5: 66.4000 (66.4000)  time: 12.2904  data: 12.2663  max mem: 2905
Test:  [10/50]  eta: 0:01:13  loss: 2.4544 (2.4948)  acc1: 48.8000 (47.4909)  acc5: 72.0000 (71.8545)  time: 1.8485  data: 1.8266  max mem: 2905
Test:  [20/50]  eta: 0:00:39  loss: 2.5602 (2.6137)  acc1: 47.2000 (46.0190)  acc5: 71.2000 (71.0476)  time: 0.7818  data: 0.7606  max mem: 2905
Test:  [30/50]  eta: 0:00:26  loss: 2.6539 (2.6283)  acc1: 46.4000 (45.9613)  acc5: 68.0000 (70.2194)  time: 1.0187  data: 0.9984  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 2.8115 (2.7271)  acc1: 44.0000 (44.3122)  acc5: 66.4000 (68.9756)  time: 1.0284  data: 1.0087  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.8396 (2.7450)  acc1: 40.8000 (44.0640)  acc5: 67.2000 (68.5600)  time: 0.5600  data: 0.5411  max mem: 2905
Test: Total time: 0:00:52 (1.0418 s / it)
* Acc@1 44.680 Acc@5 69.082 loss 2.720
Accuracy of the model on the 50000 test images: 44.7%
Max accuracy: 52.05%
Epoch: [202]  [  0/625]  eta: 3:50:43  lr: 0.001093  min_lr: 0.001093  loss: 3.0313 (3.0313)  class_acc: 0.5664 (0.5664)  weight_decay: 0.0500 (0.0500)  time: 22.1489  data: 20.6541  max mem: 2905
Epoch: [202]  [200/625]  eta: 0:14:41  lr: 0.001086  min_lr: 0.001086  loss: 2.8710 (2.8803)  class_acc: 0.5508 (0.5523)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7238 (2.9651)  time: 1.9658  data: 0.0009  max mem: 2905
Epoch: [202]  [400/625]  eta: 0:07:28  lr: 0.001080  min_lr: 0.001080  loss: 2.8897 (2.8892)  class_acc: 0.5391 (0.5506)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.0159 (3.1104)  time: 1.8130  data: 0.0007  max mem: 2905
Epoch: [202]  [600/625]  eta: 0:00:49  lr: 0.001074  min_lr: 0.001074  loss: 2.8211 (2.8919)  class_acc: 0.5547 (0.5506)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.8445 (3.1244)  time: 1.9175  data: 0.0006  max mem: 2905
Epoch: [202]  [624/625]  eta: 0:00:01  lr: 0.001073  min_lr: 0.001073  loss: 2.8772 (2.8920)  class_acc: 0.5469 (0.5504)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6209 (3.0976)  time: 0.6790  data: 0.0013  max mem: 2905
Epoch: [202] Total time: 0:20:08 (1.9335 s / it)
Averaged stats: lr: 0.001073  min_lr: 0.001073  loss: 2.8772 (2.8904)  class_acc: 0.5469 (0.5508)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6209 (3.0976)
Test:  [ 0/50]  eta: 0:09:30  loss: 2.6640 (2.6640)  acc1: 44.0000 (44.0000)  acc5: 72.8000 (72.8000)  time: 11.4009  data: 11.3727  max mem: 2905
Test:  [10/50]  eta: 0:01:11  loss: 2.4418 (2.5027)  acc1: 45.6000 (46.4727)  acc5: 72.8000 (71.5636)  time: 1.7952  data: 1.7723  max mem: 2905
Test:  [20/50]  eta: 0:00:37  loss: 2.4775 (2.5617)  acc1: 45.6000 (46.1333)  acc5: 71.2000 (70.9714)  time: 0.7547  data: 0.7341  max mem: 2905
Test:  [30/50]  eta: 0:00:21  loss: 2.5656 (2.5530)  acc1: 45.6000 (46.3742)  acc5: 70.4000 (70.8387)  time: 0.7089  data: 0.6890  max mem: 2905
Test:  [40/50]  eta: 0:00:10  loss: 2.5656 (2.5590)  acc1: 46.4000 (46.5366)  acc5: 70.4000 (70.7902)  time: 0.8996  data: 0.8797  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 2.4642 (2.5453)  acc1: 47.2000 (46.7840)  acc5: 72.0000 (71.0880)  time: 0.6991  data: 0.6796  max mem: 2905
Test: Total time: 0:00:48 (0.9610 s / it)
* Acc@1 47.558 Acc@5 71.790 loss 2.507
Accuracy of the model on the 50000 test images: 47.6%
Max accuracy: 52.05%
Epoch: [203]  [  0/625]  eta: 3:34:49  lr: 0.001073  min_lr: 0.001073  loss: 2.7870 (2.7870)  class_acc: 0.5547 (0.5547)  weight_decay: 0.0500 (0.0500)  time: 20.6225  data: 18.6708  max mem: 2905
Epoch: [203]  [200/625]  eta: 0:14:22  lr: 0.001066  min_lr: 0.001066  loss: 2.9130 (2.8775)  class_acc: 0.5469 (0.5538)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5122 (3.2295)  time: 1.9272  data: 0.1290  max mem: 2905
Epoch: [203]  [400/625]  eta: 0:07:19  lr: 0.001060  min_lr: 0.001060  loss: 2.8503 (2.8790)  class_acc: 0.5430 (0.5538)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.8870 (inf)  time: 1.9587  data: 0.1191  max mem: 2905
Epoch: [203]  [600/625]  eta: 0:00:48  lr: 0.001054  min_lr: 0.001054  loss: 2.8840 (2.8880)  class_acc: 0.5469 (0.5518)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.8902 (inf)  time: 1.9373  data: 0.0192  max mem: 2905
Epoch: [203]  [624/625]  eta: 0:00:01  lr: 0.001053  min_lr: 0.001053  loss: 2.9153 (2.8881)  class_acc: 0.5391 (0.5517)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.8295 (inf)  time: 0.4042  data: 0.0232  max mem: 2905
Epoch: [203] Total time: 0:19:49 (1.9037 s / it)
Averaged stats: lr: 0.001053  min_lr: 0.001053  loss: 2.9153 (2.8877)  class_acc: 0.5391 (0.5515)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.8295 (inf)
Test:  [ 0/50]  eta: 0:10:23  loss: 2.5038 (2.5038)  acc1: 45.6000 (45.6000)  acc5: 71.2000 (71.2000)  time: 12.4686  data: 12.4340  max mem: 2905
Test:  [10/50]  eta: 0:01:23  loss: 2.3957 (2.3556)  acc1: 48.8000 (50.0364)  acc5: 72.8000 (73.6000)  time: 2.0756  data: 2.0543  max mem: 2905
Test:  [20/50]  eta: 0:00:46  loss: 2.4901 (2.4844)  acc1: 47.2000 (47.6571)  acc5: 72.8000 (72.4571)  time: 0.9925  data: 0.9721  max mem: 2905
Test:  [30/50]  eta: 0:00:26  loss: 2.5934 (2.4566)  acc1: 48.0000 (48.8258)  acc5: 73.6000 (72.8000)  time: 0.9134  data: 0.8929  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 2.5934 (2.5026)  acc1: 48.8000 (48.0000)  acc5: 72.0000 (72.2732)  time: 0.7398  data: 0.7207  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 2.6015 (2.5182)  acc1: 44.8000 (47.5680)  acc5: 69.6000 (71.9360)  time: 0.6680  data: 0.6481  max mem: 2905
Test: Total time: 0:00:47 (0.9547 s / it)
* Acc@1 48.036 Acc@5 72.350 loss 2.484
Accuracy of the model on the 50000 test images: 48.0%
Max accuracy: 52.05%
Epoch: [204]  [  0/625]  eta: 3:38:32  lr: 0.001053  min_lr: 0.001053  loss: 2.7783 (2.7783)  class_acc: 0.5547 (0.5547)  weight_decay: 0.0500 (0.0500)  time: 20.9800  data: 20.8434  max mem: 2905
Epoch: [204]  [200/625]  eta: 0:13:57  lr: 0.001047  min_lr: 0.001047  loss: 2.8437 (2.8723)  class_acc: 0.5586 (0.5548)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.7670 (2.9545)  time: 1.9487  data: 0.7485  max mem: 2905
Epoch: [204]  [400/625]  eta: 0:07:10  lr: 0.001040  min_lr: 0.001040  loss: 2.8148 (2.8789)  class_acc: 0.5625 (0.5528)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6933 (2.9670)  time: 1.8240  data: 0.0006  max mem: 2905
Epoch: [204]  [600/625]  eta: 0:00:47  lr: 0.001034  min_lr: 0.001034  loss: 2.9004 (2.8848)  class_acc: 0.5352 (0.5519)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2490 (2.8898)  time: 1.9440  data: 0.0170  max mem: 2905
Epoch: [204]  [624/625]  eta: 0:00:01  lr: 0.001033  min_lr: 0.001033  loss: 2.8976 (2.8862)  class_acc: 0.5391 (0.5512)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2809 (2.8912)  time: 0.7434  data: 0.0018  max mem: 2905
Epoch: [204] Total time: 0:19:40 (1.8890 s / it)
Averaged stats: lr: 0.001033  min_lr: 0.001033  loss: 2.8976 (2.8854)  class_acc: 0.5391 (0.5524)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2809 (2.8912)
Test:  [ 0/50]  eta: 0:09:52  loss: 2.5162 (2.5162)  acc1: 45.6000 (45.6000)  acc5: 76.0000 (76.0000)  time: 11.8526  data: 11.8252  max mem: 2905
Test:  [10/50]  eta: 0:01:20  loss: 2.3630 (2.3850)  acc1: 48.8000 (49.7455)  acc5: 74.4000 (73.8909)  time: 2.0144  data: 1.9950  max mem: 2905
Test:  [20/50]  eta: 0:00:42  loss: 2.4599 (2.5727)  acc1: 48.8000 (46.2857)  acc5: 72.0000 (72.1143)  time: 0.9101  data: 0.8908  max mem: 2905
Test:  [30/50]  eta: 0:00:23  loss: 2.6263 (2.5818)  acc1: 45.6000 (46.4774)  acc5: 72.0000 (72.0516)  time: 0.7246  data: 0.7040  max mem: 2905
Test:  [40/50]  eta: 0:00:10  loss: 2.5700 (2.6056)  acc1: 45.6000 (46.3610)  acc5: 71.2000 (71.9220)  time: 0.6557  data: 0.6354  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 2.6339 (2.6101)  acc1: 44.0000 (46.4000)  acc5: 71.2000 (71.6640)  time: 0.4996  data: 0.4810  max mem: 2905
Test: Total time: 0:00:46 (0.9283 s / it)
* Acc@1 47.632 Acc@5 71.544 loss 2.566
Accuracy of the model on the 50000 test images: 47.6%
Max accuracy: 52.05%
Epoch: [205]  [  0/625]  eta: 3:31:44  lr: 0.001033  min_lr: 0.001033  loss: 2.6924 (2.6924)  class_acc: 0.5859 (0.5859)  weight_decay: 0.0500 (0.0500)  time: 20.3277  data: 14.7903  max mem: 2905
Epoch: [205]  [200/625]  eta: 0:14:10  lr: 0.001027  min_lr: 0.001027  loss: 2.8817 (2.8708)  class_acc: 0.5391 (0.5558)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.9875 (3.0523)  time: 1.8813  data: 0.0006  max mem: 2905
Epoch: [205]  [400/625]  eta: 0:07:20  lr: 0.001021  min_lr: 0.001021  loss: 2.9028 (2.8806)  class_acc: 0.5430 (0.5539)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5332 (3.1168)  time: 1.8082  data: 0.0008  max mem: 2905
Epoch: [205]  [600/625]  eta: 0:00:48  lr: 0.001014  min_lr: 0.001014  loss: 2.8808 (2.8818)  class_acc: 0.5508 (0.5539)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3462 (3.0097)  time: 1.8975  data: 0.0005  max mem: 2905
Epoch: [205]  [624/625]  eta: 0:00:01  lr: 0.001014  min_lr: 0.001014  loss: 2.8942 (2.8826)  class_acc: 0.5430 (0.5538)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3462 (2.9980)  time: 0.7674  data: 0.0016  max mem: 2905
Epoch: [205] Total time: 0:19:49 (1.9025 s / it)
Averaged stats: lr: 0.001014  min_lr: 0.001014  loss: 2.8942 (2.8824)  class_acc: 0.5430 (0.5529)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3462 (2.9980)
Test:  [ 0/50]  eta: 0:10:25  loss: 2.8214 (2.8214)  acc1: 41.6000 (41.6000)  acc5: 66.4000 (66.4000)  time: 12.5016  data: 12.4749  max mem: 2905
Test:  [10/50]  eta: 0:01:23  loss: 2.8214 (2.8339)  acc1: 43.2000 (41.9636)  acc5: 68.0000 (67.7818)  time: 2.0997  data: 2.0754  max mem: 2905
Test:  [20/50]  eta: 0:00:47  loss: 2.8672 (2.9050)  acc1: 41.6000 (40.7238)  acc5: 66.4000 (66.7810)  time: 1.0246  data: 1.0028  max mem: 2905
Test:  [30/50]  eta: 0:00:27  loss: 2.9260 (2.9135)  acc1: 39.2000 (40.7226)  acc5: 66.4000 (66.7355)  time: 0.9471  data: 0.9275  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 3.0545 (2.9637)  acc1: 39.2000 (40.3707)  acc5: 64.8000 (65.8732)  time: 0.7521  data: 0.7315  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 3.0545 (2.9767)  acc1: 40.8000 (40.4800)  acc5: 64.0000 (65.5520)  time: 0.7124  data: 0.6924  max mem: 2905
Test: Total time: 0:00:51 (1.0253 s / it)
* Acc@1 41.094 Acc@5 65.924 loss 2.938
Accuracy of the model on the 50000 test images: 41.1%
Max accuracy: 52.05%
Epoch: [206]  [  0/625]  eta: 3:17:04  lr: 0.001014  min_lr: 0.001014  loss: 2.8358 (2.8358)  class_acc: 0.5742 (0.5742)  weight_decay: 0.0500 (0.0500)  time: 18.9190  data: 16.1167  max mem: 2905
Epoch: [206]  [200/625]  eta: 0:13:54  lr: 0.001007  min_lr: 0.001007  loss: 2.8740 (2.8624)  class_acc: 0.5586 (0.5565)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4453 (3.1968)  time: 1.6688  data: 0.0011  max mem: 2905
Epoch: [206]  [400/625]  eta: 0:07:14  lr: 0.001001  min_lr: 0.001001  loss: 2.9083 (2.8765)  class_acc: 0.5391 (0.5551)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6642 (3.3105)  time: 1.7639  data: 0.0009  max mem: 2905
Epoch: [206]  [600/625]  eta: 0:00:48  lr: 0.000995  min_lr: 0.000995  loss: 2.9223 (2.8792)  class_acc: 0.5469 (0.5543)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0321 (3.0616)  time: 2.0394  data: 0.0008  max mem: 2905
Epoch: [206]  [624/625]  eta: 0:00:01  lr: 0.000994  min_lr: 0.000994  loss: 2.8726 (2.8799)  class_acc: 0.5508 (0.5543)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4480 (3.0424)  time: 0.7083  data: 0.0051  max mem: 2905
Epoch: [206] Total time: 0:20:02 (1.9233 s / it)
Averaged stats: lr: 0.000994  min_lr: 0.000994  loss: 2.8726 (2.8784)  class_acc: 0.5508 (0.5535)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4480 (3.0424)
Test:  [ 0/50]  eta: 0:10:41  loss: 2.2146 (2.2146)  acc1: 48.8000 (48.8000)  acc5: 80.0000 (80.0000)  time: 12.8215  data: 12.7984  max mem: 2905
Test:  [10/50]  eta: 0:01:31  loss: 2.3821 (2.4057)  acc1: 48.8000 (51.4182)  acc5: 73.6000 (73.6000)  time: 2.2761  data: 2.2567  max mem: 2905
Test:  [20/50]  eta: 0:00:54  loss: 2.6524 (2.5958)  acc1: 44.0000 (46.1333)  acc5: 71.2000 (71.2762)  time: 1.2719  data: 1.2519  max mem: 2905
Test:  [30/50]  eta: 0:00:31  loss: 2.7257 (2.6335)  acc1: 40.8000 (45.2903)  acc5: 69.6000 (70.7871)  time: 1.1572  data: 1.1369  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 2.7443 (2.6591)  acc1: 40.0000 (44.5463)  acc5: 68.0000 (70.4000)  time: 0.7079  data: 0.6886  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.7054 (2.6617)  acc1: 42.4000 (44.5600)  acc5: 69.6000 (70.3360)  time: 0.6224  data: 0.6037  max mem: 2905
Test: Total time: 0:00:53 (1.0603 s / it)
* Acc@1 45.560 Acc@5 70.562 loss 2.637
Accuracy of the model on the 50000 test images: 45.6%
Max accuracy: 52.05%
Epoch: [207]  [  0/625]  eta: 4:09:27  lr: 0.000994  min_lr: 0.000994  loss: 2.6924 (2.6924)  class_acc: 0.6172 (0.6172)  weight_decay: 0.0500 (0.0500)  time: 23.9479  data: 20.8220  max mem: 2905
Epoch: [207]  [200/625]  eta: 0:14:31  lr: 0.000988  min_lr: 0.000988  loss: 2.8869 (2.8617)  class_acc: 0.5469 (0.5564)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7443 (2.7849)  time: 2.0594  data: 0.9492  max mem: 2905
Epoch: [207]  [400/625]  eta: 0:07:35  lr: 0.000982  min_lr: 0.000982  loss: 2.8765 (2.8678)  class_acc: 0.5469 (0.5544)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3188 (2.8626)  time: 2.0815  data: 0.0008  max mem: 2905
Epoch: [207]  [600/625]  eta: 0:00:51  lr: 0.000976  min_lr: 0.000976  loss: 2.9320 (2.8744)  class_acc: 0.5430 (0.5544)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4625 (2.9755)  time: 2.2362  data: 0.0009  max mem: 2905
Epoch: [207]  [624/625]  eta: 0:00:01  lr: 0.000975  min_lr: 0.000975  loss: 2.8622 (2.8744)  class_acc: 0.5469 (0.5543)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7302 (3.0051)  time: 0.6907  data: 0.0013  max mem: 2905
Epoch: [207] Total time: 0:20:49 (1.9997 s / it)
Averaged stats: lr: 0.000975  min_lr: 0.000975  loss: 2.8622 (2.8756)  class_acc: 0.5469 (0.5546)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7302 (3.0051)
Test:  [ 0/50]  eta: 0:11:58  loss: 2.6572 (2.6572)  acc1: 39.2000 (39.2000)  acc5: 69.6000 (69.6000)  time: 14.3756  data: 14.3465  max mem: 2905
Test:  [10/50]  eta: 0:01:28  loss: 2.5540 (2.4685)  acc1: 49.6000 (47.8545)  acc5: 72.8000 (73.8909)  time: 2.2004  data: 2.1810  max mem: 2905
Test:  [20/50]  eta: 0:00:49  loss: 2.5540 (2.5972)  acc1: 47.2000 (46.6286)  acc5: 72.0000 (71.4667)  time: 1.0236  data: 1.0049  max mem: 2905
Test:  [30/50]  eta: 0:00:29  loss: 2.4939 (2.5700)  acc1: 47.2000 (47.2258)  acc5: 70.4000 (71.9742)  time: 1.0658  data: 1.0470  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 2.6372 (2.6029)  acc1: 46.4000 (46.6927)  acc5: 71.2000 (71.3366)  time: 0.9092  data: 0.8903  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.6372 (2.6001)  acc1: 46.4000 (46.8960)  acc5: 71.2000 (71.3600)  time: 0.7616  data: 0.7423  max mem: 2905
Test: Total time: 0:00:56 (1.1338 s / it)
* Acc@1 47.810 Acc@5 71.930 loss 2.556
Accuracy of the model on the 50000 test images: 47.8%
Max accuracy: 52.05%
Epoch: [208]  [  0/625]  eta: 3:56:25  lr: 0.000975  min_lr: 0.000975  loss: 3.0261 (3.0261)  class_acc: 0.4922 (0.4922)  weight_decay: 0.0500 (0.0500)  time: 22.6962  data: 21.7468  max mem: 2905
Epoch: [208]  [200/625]  eta: 0:14:58  lr: 0.000969  min_lr: 0.000969  loss: 2.8062 (2.8636)  class_acc: 0.5664 (0.5572)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1851 (3.0015)  time: 1.8681  data: 0.0548  max mem: 2905
Epoch: [208]  [400/625]  eta: 0:07:38  lr: 0.000963  min_lr: 0.000963  loss: 2.8370 (2.8658)  class_acc: 0.5547 (0.5562)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6270 (3.0561)  time: 1.9013  data: 0.0008  max mem: 2905
Epoch: [208]  [600/625]  eta: 0:00:50  lr: 0.000956  min_lr: 0.000956  loss: 2.8803 (2.8694)  class_acc: 0.5508 (0.5554)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.1440 (3.0891)  time: 2.0358  data: 0.0008  max mem: 2905
Epoch: [208]  [624/625]  eta: 0:00:01  lr: 0.000956  min_lr: 0.000956  loss: 2.8772 (2.8712)  class_acc: 0.5430 (0.5550)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.8043 (3.0737)  time: 0.8078  data: 0.0017  max mem: 2905
Epoch: [208] Total time: 0:20:31 (1.9697 s / it)
Averaged stats: lr: 0.000956  min_lr: 0.000956  loss: 2.8772 (2.8711)  class_acc: 0.5430 (0.5555)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.8043 (3.0737)
Test:  [ 0/50]  eta: 0:10:05  loss: 2.2059 (2.2059)  acc1: 54.4000 (54.4000)  acc5: 76.8000 (76.8000)  time: 12.1020  data: 12.0707  max mem: 2905
Test:  [10/50]  eta: 0:01:19  loss: 1.9689 (2.0178)  acc1: 56.8000 (56.6545)  acc5: 79.2000 (79.9273)  time: 1.9765  data: 1.9570  max mem: 2905
Test:  [20/50]  eta: 0:00:44  loss: 2.1178 (2.1763)  acc1: 52.8000 (53.6762)  acc5: 78.4000 (78.4381)  time: 0.9524  data: 0.9325  max mem: 2905
Test:  [30/50]  eta: 0:00:25  loss: 2.1953 (2.2158)  acc1: 50.4000 (52.8516)  acc5: 76.8000 (77.7290)  time: 0.9173  data: 0.8972  max mem: 2905
Test:  [40/50]  eta: 0:00:10  loss: 2.1433 (2.2386)  acc1: 50.4000 (52.7610)  acc5: 76.0000 (77.5220)  time: 0.6740  data: 0.6556  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 2.1799 (2.2359)  acc1: 50.4000 (52.4640)  acc5: 76.8000 (77.6000)  time: 0.5923  data: 0.5743  max mem: 2905
Test: Total time: 0:00:46 (0.9310 s / it)
* Acc@1 53.338 Acc@5 77.504 loss 2.205
Accuracy of the model on the 50000 test images: 53.3%
Max accuracy: 53.34%
Epoch: [209]  [  0/625]  eta: 3:27:12  lr: 0.000956  min_lr: 0.000956  loss: 2.9900 (2.9900)  class_acc: 0.5352 (0.5352)  weight_decay: 0.0500 (0.0500)  time: 19.8919  data: 19.7616  max mem: 2905
Epoch: [209]  [200/625]  eta: 0:13:49  lr: 0.000950  min_lr: 0.000950  loss: 2.8705 (2.8713)  class_acc: 0.5430 (0.5553)  weight_decay: 0.0500 (0.0500)  grad_norm: 5.3777 (3.5759)  time: 1.7230  data: 0.0007  max mem: 2905
Epoch: [209]  [400/625]  eta: 0:07:14  lr: 0.000944  min_lr: 0.000944  loss: 2.8353 (2.8723)  class_acc: 0.5508 (0.5546)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6873 (3.4149)  time: 1.9223  data: 0.1811  max mem: 2905
Epoch: [209]  [600/625]  eta: 0:00:48  lr: 0.000937  min_lr: 0.000937  loss: 2.8240 (2.8762)  class_acc: 0.5664 (0.5539)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4924 (3.4393)  time: 2.0451  data: 0.0374  max mem: 2905
Epoch: [209]  [624/625]  eta: 0:00:01  lr: 0.000937  min_lr: 0.000937  loss: 2.8606 (2.8772)  class_acc: 0.5469 (0.5537)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4966 (3.4208)  time: 0.9492  data: 0.0124  max mem: 2905
Epoch: [209] Total time: 0:19:42 (1.8918 s / it)
Averaged stats: lr: 0.000937  min_lr: 0.000937  loss: 2.8606 (2.8691)  class_acc: 0.5469 (0.5564)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4966 (3.4208)
Test:  [ 0/50]  eta: 0:10:40  loss: 2.4309 (2.4309)  acc1: 46.4000 (46.4000)  acc5: 71.2000 (71.2000)  time: 12.8180  data: 12.7911  max mem: 2905
Test:  [10/50]  eta: 0:01:21  loss: 2.3890 (2.3352)  acc1: 50.4000 (50.5455)  acc5: 74.4000 (73.7455)  time: 2.0322  data: 2.0108  max mem: 2905
Test:  [20/50]  eta: 0:00:47  loss: 2.4491 (2.5133)  acc1: 46.4000 (46.8571)  acc5: 73.6000 (72.1524)  time: 1.0141  data: 0.9944  max mem: 2905
Test:  [30/50]  eta: 0:00:28  loss: 2.5529 (2.5310)  acc1: 43.2000 (46.8903)  acc5: 71.2000 (71.8452)  time: 1.0555  data: 1.0372  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 2.5512 (2.5667)  acc1: 45.6000 (47.0439)  acc5: 71.2000 (71.3171)  time: 0.8003  data: 0.7822  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.5972 (2.5768)  acc1: 47.2000 (47.0400)  acc5: 68.0000 (70.9760)  time: 0.6907  data: 0.6723  max mem: 2905
Test: Total time: 0:00:51 (1.0250 s / it)
* Acc@1 47.846 Acc@5 71.524 loss 2.530
Accuracy of the model on the 50000 test images: 47.8%
Max accuracy: 53.34%
Epoch: [210]  [  0/625]  eta: 3:14:37  lr: 0.000937  min_lr: 0.000937  loss: 2.8904 (2.8904)  class_acc: 0.5781 (0.5781)  weight_decay: 0.0500 (0.0500)  time: 18.6843  data: 17.7713  max mem: 2905
Epoch: [210]  [200/625]  eta: 0:14:23  lr: 0.000931  min_lr: 0.000931  loss: 2.8605 (2.8540)  class_acc: 0.5586 (0.5611)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7377 (2.7882)  time: 1.8635  data: 0.0008  max mem: 2905
Epoch: [210]  [400/625]  eta: 0:07:19  lr: 0.000925  min_lr: 0.000925  loss: 2.8761 (2.8583)  class_acc: 0.5469 (0.5586)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.1565 (2.8829)  time: 1.8057  data: 0.0007  max mem: 2905
Epoch: [210]  [600/625]  eta: 0:00:48  lr: 0.000918  min_lr: 0.000918  loss: 2.8734 (2.8645)  class_acc: 0.5625 (0.5572)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.7189 (2.9967)  time: 2.0221  data: 0.0009  max mem: 2905
Epoch: [210]  [624/625]  eta: 0:00:01  lr: 0.000918  min_lr: 0.000918  loss: 2.8697 (2.8653)  class_acc: 0.5430 (0.5568)  weight_decay: 0.0500 (0.0500)  grad_norm: 4.3915 (3.0620)  time: 0.7600  data: 0.0012  max mem: 2905
Epoch: [210] Total time: 0:19:45 (1.8974 s / it)
Averaged stats: lr: 0.000918  min_lr: 0.000918  loss: 2.8697 (2.8671)  class_acc: 0.5430 (0.5565)  weight_decay: 0.0500 (0.0500)  grad_norm: 4.3915 (3.0620)
Test:  [ 0/50]  eta: 0:09:13  loss: 1.8271 (1.8271)  acc1: 55.2000 (55.2000)  acc5: 84.0000 (84.0000)  time: 11.0732  data: 11.0440  max mem: 2905
Test:  [10/50]  eta: 0:01:22  loss: 1.9140 (1.9644)  acc1: 58.4000 (58.6909)  acc5: 81.6000 (80.2182)  time: 2.0627  data: 2.0433  max mem: 2905
Test:  [20/50]  eta: 0:00:50  loss: 2.1921 (2.1230)  acc1: 52.0000 (54.5524)  acc5: 77.6000 (78.2095)  time: 1.2114  data: 1.1929  max mem: 2905
Test:  [30/50]  eta: 0:00:29  loss: 2.2073 (2.1345)  acc1: 52.0000 (54.8387)  acc5: 76.0000 (77.9097)  time: 1.1808  data: 1.1616  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 2.2189 (2.1653)  acc1: 52.8000 (54.3024)  acc5: 76.8000 (77.3659)  time: 0.7509  data: 0.7302  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.2722 (2.1755)  acc1: 51.2000 (53.6800)  acc5: 76.8000 (77.4080)  time: 0.6409  data: 0.6211  max mem: 2905
Test: Total time: 0:00:51 (1.0256 s / it)
* Acc@1 54.420 Acc@5 78.240 loss 2.133
Accuracy of the model on the 50000 test images: 54.4%
Max accuracy: 54.42%
Epoch: [211]  [  0/625]  eta: 3:26:31  lr: 0.000918  min_lr: 0.000918  loss: 2.7751 (2.7751)  class_acc: 0.5430 (0.5430)  weight_decay: 0.0500 (0.0500)  time: 19.8269  data: 17.6656  max mem: 2905
Epoch: [211]  [200/625]  eta: 0:13:29  lr: 0.000912  min_lr: 0.000912  loss: 2.8535 (2.8510)  class_acc: 0.5469 (0.5608)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.8003 (3.1526)  time: 1.8632  data: 0.0007  max mem: 2905
Epoch: [211]  [400/625]  eta: 0:07:02  lr: 0.000906  min_lr: 0.000906  loss: 2.8203 (2.8542)  class_acc: 0.5625 (0.5593)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7219 (3.1742)  time: 1.9107  data: 0.0007  max mem: 2905
Epoch: [211]  [600/625]  eta: 0:00:47  lr: 0.000900  min_lr: 0.000900  loss: 2.9224 (2.8626)  class_acc: 0.5508 (0.5574)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.3338 (3.2063)  time: 2.0623  data: 0.1461  max mem: 2905
Epoch: [211]  [624/625]  eta: 0:00:01  lr: 0.000899  min_lr: 0.000899  loss: 2.8189 (2.8629)  class_acc: 0.5586 (0.5574)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.4208 (3.2116)  time: 0.8514  data: 0.2455  max mem: 2905
Epoch: [211] Total time: 0:19:18 (1.8535 s / it)
Averaged stats: lr: 0.000899  min_lr: 0.000899  loss: 2.8189 (2.8648)  class_acc: 0.5586 (0.5573)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.4208 (3.2116)
Test:  [ 0/50]  eta: 0:09:22  loss: 2.6556 (2.6556)  acc1: 41.6000 (41.6000)  acc5: 69.6000 (69.6000)  time: 11.2446  data: 11.2208  max mem: 2905
Test:  [10/50]  eta: 0:01:17  loss: 2.4608 (2.4689)  acc1: 47.2000 (48.0000)  acc5: 72.0000 (72.5818)  time: 1.9351  data: 1.9150  max mem: 2905
Test:  [20/50]  eta: 0:00:45  loss: 2.6005 (2.6847)  acc1: 43.2000 (43.7714)  acc5: 68.8000 (70.2095)  time: 1.0431  data: 1.0233  max mem: 2905
Test:  [30/50]  eta: 0:00:26  loss: 2.7997 (2.6992)  acc1: 40.8000 (43.7161)  acc5: 67.2000 (69.0839)  time: 0.9903  data: 0.9702  max mem: 2905
Test:  [40/50]  eta: 0:00:10  loss: 2.7758 (2.7548)  acc1: 43.2000 (43.4732)  acc5: 67.2000 (68.5268)  time: 0.6178  data: 0.5963  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 2.7928 (2.7611)  acc1: 42.4000 (43.4560)  acc5: 64.8000 (68.2240)  time: 0.5405  data: 0.5197  max mem: 2905
Test: Total time: 0:00:45 (0.9039 s / it)
* Acc@1 44.112 Acc@5 69.138 loss 2.721
Accuracy of the model on the 50000 test images: 44.1%
Max accuracy: 54.42%
Epoch: [212]  [  0/625]  eta: 4:36:50  lr: 0.000899  min_lr: 0.000899  loss: 2.7424 (2.7424)  class_acc: 0.5625 (0.5625)  weight_decay: 0.0500 (0.0500)  time: 26.5775  data: 17.8196  max mem: 2905
Epoch: [212]  [200/625]  eta: 0:13:50  lr: 0.000893  min_lr: 0.000893  loss: 2.9033 (2.8590)  class_acc: 0.5586 (0.5614)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6885 (2.6922)  time: 1.7908  data: 0.0008  max mem: 2905
Epoch: [212]  [400/625]  eta: 0:07:02  lr: 0.000887  min_lr: 0.000887  loss: 2.9157 (2.8567)  class_acc: 0.5508 (0.5599)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6200 (2.9315)  time: 1.8920  data: 0.0006  max mem: 2905
Epoch: [212]  [600/625]  eta: 0:00:46  lr: 0.000881  min_lr: 0.000881  loss: 2.9140 (2.8588)  class_acc: 0.5547 (0.5582)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.9470 (2.9191)  time: 1.8899  data: 0.0007  max mem: 2905
Epoch: [212]  [624/625]  eta: 0:00:01  lr: 0.000880  min_lr: 0.000880  loss: 2.8315 (2.8584)  class_acc: 0.5508 (0.5581)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7213 (2.9196)  time: 0.8123  data: 0.0020  max mem: 2905
Epoch: [212] Total time: 0:19:15 (1.8490 s / it)
Averaged stats: lr: 0.000880  min_lr: 0.000880  loss: 2.8315 (2.8586)  class_acc: 0.5508 (0.5583)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7213 (2.9196)
Test:  [ 0/50]  eta: 0:09:58  loss: 2.9251 (2.9251)  acc1: 38.4000 (38.4000)  acc5: 62.4000 (62.4000)  time: 11.9609  data: 11.9297  max mem: 2905
Test:  [10/50]  eta: 0:01:24  loss: 2.4204 (2.3933)  acc1: 51.2000 (49.9636)  acc5: 73.6000 (72.7273)  time: 2.1246  data: 2.1054  max mem: 2905
Test:  [20/50]  eta: 0:00:50  loss: 2.6036 (2.6218)  acc1: 45.6000 (46.0571)  acc5: 69.6000 (69.9810)  time: 1.1581  data: 1.1389  max mem: 2905
Test:  [30/50]  eta: 0:00:29  loss: 2.7284 (2.6017)  acc1: 44.0000 (46.7097)  acc5: 68.0000 (70.1936)  time: 1.1218  data: 1.1021  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 2.6970 (2.6432)  acc1: 46.4000 (46.2049)  acc5: 68.0000 (69.7561)  time: 0.7653  data: 0.7451  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.7736 (2.6540)  acc1: 45.6000 (45.8720)  acc5: 68.0000 (69.6480)  time: 0.6503  data: 0.6277  max mem: 2905
Test: Total time: 0:00:51 (1.0223 s / it)
* Acc@1 46.210 Acc@5 70.362 loss 2.627
Accuracy of the model on the 50000 test images: 46.2%
Max accuracy: 54.42%
Epoch: [213]  [  0/625]  eta: 3:24:59  lr: 0.000880  min_lr: 0.000880  loss: 2.7151 (2.7151)  class_acc: 0.5430 (0.5430)  weight_decay: 0.0500 (0.0500)  time: 19.6793  data: 19.5196  max mem: 2905
Epoch: [213]  [200/625]  eta: 0:14:13  lr: 0.000874  min_lr: 0.000874  loss: 2.8313 (2.8353)  class_acc: 0.5547 (0.5618)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7075 (inf)  time: 1.9866  data: 0.0007  max mem: 2905
Epoch: [213]  [400/625]  eta: 0:07:16  lr: 0.000868  min_lr: 0.000868  loss: 2.8347 (2.8456)  class_acc: 0.5703 (0.5601)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.2377 (inf)  time: 1.8750  data: 0.4142  max mem: 2905
Epoch: [213]  [600/625]  eta: 0:00:48  lr: 0.000863  min_lr: 0.000863  loss: 2.8553 (2.8525)  class_acc: 0.5547 (0.5593)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2295 (inf)  time: 2.1345  data: 0.0007  max mem: 2905
Epoch: [213]  [624/625]  eta: 0:00:01  lr: 0.000862  min_lr: 0.000862  loss: 2.8048 (2.8519)  class_acc: 0.5625 (0.5595)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.2696 (inf)  time: 0.8310  data: 0.0340  max mem: 2905
Epoch: [213] Total time: 0:19:46 (1.8984 s / it)
Averaged stats: lr: 0.000862  min_lr: 0.000862  loss: 2.8048 (2.8573)  class_acc: 0.5625 (0.5587)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.2696 (inf)
Test:  [ 0/50]  eta: 0:09:41  loss: 2.2315 (2.2315)  acc1: 51.2000 (51.2000)  acc5: 75.2000 (75.2000)  time: 11.6324  data: 11.6045  max mem: 2905
Test:  [10/50]  eta: 0:01:32  loss: 1.8380 (1.8331)  acc1: 61.6000 (59.1273)  acc5: 80.8000 (80.4364)  time: 2.3106  data: 2.2897  max mem: 2905
Test:  [20/50]  eta: 0:00:54  loss: 1.9574 (2.0150)  acc1: 56.0000 (55.1238)  acc5: 79.2000 (79.2000)  time: 1.3276  data: 1.3074  max mem: 2905
Test:  [30/50]  eta: 0:00:31  loss: 2.1893 (2.0638)  acc1: 51.2000 (54.1677)  acc5: 76.8000 (78.2452)  time: 1.1706  data: 1.1510  max mem: 2905
Test:  [40/50]  eta: 0:00:13  loss: 2.1573 (2.0808)  acc1: 51.2000 (53.8927)  acc5: 76.8000 (78.2634)  time: 0.8058  data: 0.7857  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.1114 (2.0973)  acc1: 53.6000 (53.7440)  acc5: 79.2000 (78.3840)  time: 0.7066  data: 0.6858  max mem: 2905
Test: Total time: 0:00:56 (1.1298 s / it)
* Acc@1 54.930 Acc@5 79.038 loss 2.047
Accuracy of the model on the 50000 test images: 54.9%
Max accuracy: 54.93%
Epoch: [214]  [  0/625]  eta: 3:58:26  lr: 0.000862  min_lr: 0.000862  loss: 2.7778 (2.7778)  class_acc: 0.5742 (0.5742)  weight_decay: 0.0500 (0.0500)  time: 22.8907  data: 20.8083  max mem: 2905
Epoch: [214]  [200/625]  eta: 0:14:42  lr: 0.000856  min_lr: 0.000856  loss: 2.8224 (2.8515)  class_acc: 0.5586 (0.5594)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7554 (3.1999)  time: 1.9666  data: 1.7635  max mem: 2905
Epoch: [214]  [400/625]  eta: 0:07:38  lr: 0.000850  min_lr: 0.000850  loss: 2.8676 (2.8529)  class_acc: 0.5625 (0.5593)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.2494 (3.0852)  time: 1.9234  data: 0.0009  max mem: 2905
Epoch: [214]  [600/625]  eta: 0:00:50  lr: 0.000844  min_lr: 0.000844  loss: 2.8551 (2.8546)  class_acc: 0.5547 (0.5593)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.8018 (3.2174)  time: 1.9554  data: 0.0437  max mem: 2905
Epoch: [214]  [624/625]  eta: 0:00:01  lr: 0.000844  min_lr: 0.000844  loss: 2.8616 (2.8548)  class_acc: 0.5508 (0.5592)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.4322 (3.2245)  time: 0.8874  data: 0.0297  max mem: 2905
Epoch: [214] Total time: 0:20:36 (1.9784 s / it)
Averaged stats: lr: 0.000844  min_lr: 0.000844  loss: 2.8616 (2.8526)  class_acc: 0.5508 (0.5597)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.4322 (3.2245)
Test:  [ 0/50]  eta: 0:10:27  loss: 1.8443 (1.8443)  acc1: 59.2000 (59.2000)  acc5: 84.8000 (84.8000)  time: 12.5563  data: 12.5030  max mem: 2905
Test:  [10/50]  eta: 0:01:26  loss: 1.8811 (1.9989)  acc1: 58.4000 (56.5091)  acc5: 82.4000 (80.5091)  time: 2.1748  data: 2.1497  max mem: 2905
Test:  [20/50]  eta: 0:00:51  loss: 2.2505 (2.2000)  acc1: 52.0000 (53.4857)  acc5: 76.8000 (77.8286)  time: 1.1660  data: 1.1448  max mem: 2905
Test:  [30/50]  eta: 0:00:29  loss: 2.2456 (2.2004)  acc1: 49.6000 (52.8516)  acc5: 74.4000 (77.2645)  time: 1.0801  data: 1.0607  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 2.1910 (2.2062)  acc1: 50.4000 (53.0732)  acc5: 75.2000 (76.9171)  time: 0.6585  data: 0.6403  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.1012 (2.1918)  acc1: 52.0000 (52.9920)  acc5: 76.8000 (77.1360)  time: 0.5313  data: 0.5119  max mem: 2905
Test: Total time: 0:00:50 (1.0048 s / it)
* Acc@1 53.612 Acc@5 77.116 loss 2.167
Accuracy of the model on the 50000 test images: 53.6%
Max accuracy: 54.93%
Epoch: [215]  [  0/625]  eta: 3:40:29  lr: 0.000843  min_lr: 0.000843  loss: 2.8950 (2.8950)  class_acc: 0.5781 (0.5781)  weight_decay: 0.0500 (0.0500)  time: 21.1671  data: 16.5604  max mem: 2905
Epoch: [215]  [200/625]  eta: 0:14:20  lr: 0.000838  min_lr: 0.000838  loss: 2.7813 (2.8414)  class_acc: 0.5586 (0.5603)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.6882 (3.5993)  time: 1.8034  data: 0.1104  max mem: 2905
Epoch: [215]  [400/625]  eta: 0:07:16  lr: 0.000832  min_lr: 0.000832  loss: 2.8371 (2.8471)  class_acc: 0.5586 (0.5594)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.0763 (3.4911)  time: 1.8987  data: 0.5416  max mem: 2905
Epoch: [215]  [600/625]  eta: 0:00:48  lr: 0.000826  min_lr: 0.000826  loss: 2.8327 (2.8493)  class_acc: 0.5508 (0.5585)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0044 (3.2370)  time: 1.9963  data: 0.5071  max mem: 2905
Epoch: [215]  [624/625]  eta: 0:00:01  lr: 0.000825  min_lr: 0.000825  loss: 2.8587 (2.8496)  class_acc: 0.5625 (0.5585)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.4475 (3.2905)  time: 0.8059  data: 0.0362  max mem: 2905
Epoch: [215] Total time: 0:19:47 (1.8998 s / it)
Averaged stats: lr: 0.000825  min_lr: 0.000825  loss: 2.8587 (2.8514)  class_acc: 0.5625 (0.5602)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.4475 (3.2905)
Test:  [ 0/50]  eta: 0:10:27  loss: 2.0447 (2.0447)  acc1: 52.0000 (52.0000)  acc5: 80.0000 (80.0000)  time: 12.5470  data: 12.5222  max mem: 2905
Test:  [10/50]  eta: 0:01:18  loss: 2.0447 (2.0176)  acc1: 55.2000 (55.4909)  acc5: 80.0000 (78.9091)  time: 1.9625  data: 1.9415  max mem: 2905
Test:  [20/50]  eta: 0:00:44  loss: 2.1533 (2.2046)  acc1: 50.4000 (51.8476)  acc5: 76.8000 (76.8000)  time: 0.9194  data: 0.8990  max mem: 2905
Test:  [30/50]  eta: 0:00:24  loss: 2.3237 (2.2326)  acc1: 49.6000 (51.5355)  acc5: 74.4000 (76.3097)  time: 0.8371  data: 0.8167  max mem: 2905
Test:  [40/50]  eta: 0:00:10  loss: 2.2201 (2.2387)  acc1: 49.6000 (51.4732)  acc5: 76.0000 (76.1951)  time: 0.5905  data: 0.5692  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 2.2400 (2.2682)  acc1: 49.6000 (51.0720)  acc5: 75.2000 (75.5360)  time: 0.5230  data: 0.5001  max mem: 2905
Test: Total time: 0:00:45 (0.9122 s / it)
* Acc@1 52.106 Acc@5 76.104 loss 2.221
Accuracy of the model on the 50000 test images: 52.1%
Max accuracy: 54.93%
Epoch: [216]  [  0/625]  eta: 3:40:31  lr: 0.000825  min_lr: 0.000825  loss: 3.0004 (3.0004)  class_acc: 0.5195 (0.5195)  weight_decay: 0.0500 (0.0500)  time: 21.1708  data: 16.5791  max mem: 2905
Epoch: [216]  [200/625]  eta: 0:14:06  lr: 0.000819  min_lr: 0.000819  loss: 2.8070 (2.8458)  class_acc: 0.5508 (0.5618)  weight_decay: 0.0500 (0.0500)  grad_norm: 4.0907 (3.2868)  time: 1.9264  data: 0.0056  max mem: 2905
Epoch: [216]  [400/625]  eta: 0:07:26  lr: 0.000814  min_lr: 0.000814  loss: 2.8496 (2.8463)  class_acc: 0.5742 (0.5619)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3506 (3.1835)  time: 2.1112  data: 0.0007  max mem: 2905
Epoch: [216]  [600/625]  eta: 0:00:49  lr: 0.000808  min_lr: 0.000808  loss: 2.8175 (2.8458)  class_acc: 0.5547 (0.5622)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6576 (3.0471)  time: 2.1839  data: 0.0007  max mem: 2905
Epoch: [216]  [624/625]  eta: 0:00:01  lr: 0.000807  min_lr: 0.000807  loss: 2.8191 (2.8459)  class_acc: 0.5664 (0.5621)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5842 (3.0354)  time: 1.0618  data: 0.0014  max mem: 2905
Epoch: [216] Total time: 0:20:18 (1.9493 s / it)
Averaged stats: lr: 0.000807  min_lr: 0.000807  loss: 2.8191 (2.8505)  class_acc: 0.5664 (0.5607)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5842 (3.0354)
Test:  [ 0/50]  eta: 0:10:39  loss: 2.2310 (2.2310)  acc1: 47.2000 (47.2000)  acc5: 75.2000 (75.2000)  time: 12.7974  data: 12.7611  max mem: 2905
Test:  [10/50]  eta: 0:01:27  loss: 1.8731 (1.9213)  acc1: 57.6000 (57.3091)  acc5: 80.8000 (80.2909)  time: 2.1791  data: 2.1569  max mem: 2905
Test:  [20/50]  eta: 0:00:52  loss: 2.0377 (2.0957)  acc1: 53.6000 (53.5619)  acc5: 77.6000 (78.1714)  time: 1.1921  data: 1.1717  max mem: 2905
Test:  [30/50]  eta: 0:00:31  loss: 2.1861 (2.1070)  acc1: 52.8000 (53.6774)  acc5: 75.2000 (77.8065)  time: 1.2187  data: 1.1984  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 2.0831 (2.1169)  acc1: 52.8000 (53.6976)  acc5: 77.6000 (77.7366)  time: 0.8201  data: 0.7998  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.0815 (2.1230)  acc1: 52.8000 (53.4240)  acc5: 78.4000 (77.9200)  time: 0.7780  data: 0.7587  max mem: 2905
Test: Total time: 0:00:53 (1.0745 s / it)
* Acc@1 54.462 Acc@5 78.244 loss 2.088
Accuracy of the model on the 50000 test images: 54.5%
Max accuracy: 54.93%
Epoch: [217]  [  0/625]  eta: 3:53:24  lr: 0.000807  min_lr: 0.000807  loss: 2.9204 (2.9204)  class_acc: 0.5820 (0.5820)  weight_decay: 0.0500 (0.0500)  time: 22.4080  data: 17.3745  max mem: 2905
Epoch: [217]  [200/625]  eta: 0:14:10  lr: 0.000801  min_lr: 0.000801  loss: 2.8442 (2.8423)  class_acc: 0.5586 (0.5623)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.1329 (3.0758)  time: 1.9887  data: 0.0007  max mem: 2905
Epoch: [217]  [400/625]  eta: 0:07:18  lr: 0.000796  min_lr: 0.000796  loss: 2.8225 (2.8415)  class_acc: 0.5547 (0.5631)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.5125 (3.3279)  time: 1.8608  data: 0.0007  max mem: 2905
Epoch: [217]  [600/625]  eta: 0:00:49  lr: 0.000790  min_lr: 0.000790  loss: 2.8185 (2.8459)  class_acc: 0.5547 (0.5613)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7362 (3.2111)  time: 2.0081  data: 0.0008  max mem: 2905
Epoch: [217]  [624/625]  eta: 0:00:01  lr: 0.000789  min_lr: 0.000789  loss: 2.7639 (2.8446)  class_acc: 0.5820 (0.5618)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1391 (3.1896)  time: 0.8480  data: 0.0017  max mem: 2905
Epoch: [217] Total time: 0:20:01 (1.9220 s / it)
Averaged stats: lr: 0.000789  min_lr: 0.000789  loss: 2.7639 (2.8447)  class_acc: 0.5820 (0.5620)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1391 (3.1896)
Test:  [ 0/50]  eta: 0:10:44  loss: 2.9077 (2.9077)  acc1: 40.8000 (40.8000)  acc5: 62.4000 (62.4000)  time: 12.8834  data: 12.8567  max mem: 2905
Test:  [10/50]  eta: 0:01:23  loss: 2.6843 (2.6003)  acc1: 47.2000 (44.9455)  acc5: 69.6000 (69.0909)  time: 2.0829  data: 2.0624  max mem: 2905
Test:  [20/50]  eta: 0:00:47  loss: 2.7803 (2.7979)  acc1: 41.6000 (42.4000)  acc5: 67.2000 (66.8571)  time: 1.0067  data: 0.9873  max mem: 2905
Test:  [30/50]  eta: 0:00:27  loss: 2.9680 (2.8100)  acc1: 39.2000 (42.4000)  acc5: 65.6000 (66.8129)  time: 0.9720  data: 0.9532  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 2.8448 (2.8293)  acc1: 39.2000 (42.2829)  acc5: 66.4000 (67.1610)  time: 0.7156  data: 0.6958  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 2.9717 (2.8423)  acc1: 41.6000 (42.2560)  acc5: 66.4000 (67.0080)  time: 0.5978  data: 0.5772  max mem: 2905
Test: Total time: 0:00:49 (0.9960 s / it)
* Acc@1 42.578 Acc@5 67.498 loss 2.796
Accuracy of the model on the 50000 test images: 42.6%
Max accuracy: 54.93%
Epoch: [218]  [  0/625]  eta: 3:51:37  lr: 0.000789  min_lr: 0.000789  loss: 2.8966 (2.8966)  class_acc: 0.5469 (0.5469)  weight_decay: 0.0500 (0.0500)  time: 22.2367  data: 17.6602  max mem: 2905
Epoch: [218]  [200/625]  eta: 0:14:32  lr: 0.000784  min_lr: 0.000784  loss: 2.8269 (2.8308)  class_acc: 0.5664 (0.5644)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1742 (3.6475)  time: 1.9986  data: 0.0355  max mem: 2905
Epoch: [218]  [400/625]  eta: 0:07:24  lr: 0.000778  min_lr: 0.000778  loss: 2.8209 (2.8363)  class_acc: 0.5586 (0.5632)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4928 (3.5446)  time: 1.9136  data: 0.0086  max mem: 2905
Epoch: [218]  [600/625]  eta: 0:00:49  lr: 0.000772  min_lr: 0.000772  loss: 2.8345 (2.8411)  class_acc: 0.5547 (0.5610)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4417 (3.4768)  time: 2.1508  data: 0.0043  max mem: 2905
Epoch: [218]  [624/625]  eta: 0:00:01  lr: 0.000772  min_lr: 0.000772  loss: 2.8303 (2.8411)  class_acc: 0.5547 (0.5610)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4417 (3.4883)  time: 0.7725  data: 0.0016  max mem: 2905
Epoch: [218] Total time: 0:20:11 (1.9385 s / it)
Averaged stats: lr: 0.000772  min_lr: 0.000772  loss: 2.8303 (2.8420)  class_acc: 0.5547 (0.5619)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4417 (3.4883)
Test:  [ 0/50]  eta: 0:12:11  loss: 2.1146 (2.1146)  acc1: 52.8000 (52.8000)  acc5: 80.8000 (80.8000)  time: 14.6324  data: 14.5998  max mem: 2905
Test:  [10/50]  eta: 0:01:28  loss: 1.9208 (1.9196)  acc1: 56.0000 (57.6727)  acc5: 80.8000 (80.7273)  time: 2.2126  data: 2.1911  max mem: 2905
Test:  [20/50]  eta: 0:00:50  loss: 2.0204 (2.1221)  acc1: 55.2000 (55.3524)  acc5: 78.4000 (78.6286)  time: 1.0435  data: 1.0239  max mem: 2905
Test:  [30/50]  eta: 0:00:29  loss: 2.2861 (2.1400)  acc1: 53.6000 (54.6065)  acc5: 76.0000 (78.1419)  time: 1.0841  data: 1.0650  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 2.0928 (2.1398)  acc1: 52.0000 (54.3024)  acc5: 76.0000 (77.5805)  time: 0.8826  data: 0.8625  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.0954 (2.1521)  acc1: 52.0000 (54.2240)  acc5: 77.6000 (77.4080)  time: 0.8566  data: 0.8348  max mem: 2905
Test: Total time: 0:00:53 (1.0720 s / it)
* Acc@1 55.472 Acc@5 78.604 loss 2.097
Accuracy of the model on the 50000 test images: 55.5%
Max accuracy: 55.47%
Epoch: [219]  [  0/625]  eta: 3:25:59  lr: 0.000771  min_lr: 0.000771  loss: 2.8030 (2.8030)  class_acc: 0.5664 (0.5664)  weight_decay: 0.0500 (0.0500)  time: 19.7753  data: 18.4014  max mem: 2905
Epoch: [219]  [200/625]  eta: 0:14:27  lr: 0.000766  min_lr: 0.000766  loss: 2.7904 (2.8293)  class_acc: 0.5664 (0.5654)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4579 (3.1716)  time: 2.0906  data: 0.3221  max mem: 2905
Epoch: [219]  [400/625]  eta: 0:07:31  lr: 0.000760  min_lr: 0.000760  loss: 2.8437 (2.8308)  class_acc: 0.5586 (0.5645)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3487 (2.9924)  time: 1.9062  data: 0.0012  max mem: 2905
Epoch: [219]  [600/625]  eta: 0:00:50  lr: 0.000755  min_lr: 0.000755  loss: 2.8752 (2.8355)  class_acc: 0.5586 (0.5638)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6165 (3.0341)  time: 2.0310  data: 0.0007  max mem: 2905
Epoch: [219]  [624/625]  eta: 0:00:01  lr: 0.000754  min_lr: 0.000754  loss: 2.8732 (2.8374)  class_acc: 0.5508 (0.5633)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2060 (3.0129)  time: 1.1016  data: 0.0078  max mem: 2905
Epoch: [219] Total time: 0:20:32 (1.9718 s / it)
Averaged stats: lr: 0.000754  min_lr: 0.000754  loss: 2.8732 (2.8364)  class_acc: 0.5508 (0.5634)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2060 (3.0129)
Test:  [ 0/50]  eta: 0:10:31  loss: 2.1764 (2.1764)  acc1: 48.8000 (48.8000)  acc5: 79.2000 (79.2000)  time: 12.6317  data: 12.5952  max mem: 2905
Test:  [10/50]  eta: 0:01:30  loss: 1.9786 (2.0772)  acc1: 56.8000 (55.0545)  acc5: 77.6000 (77.6727)  time: 2.2674  data: 2.2435  max mem: 2905
Test:  [20/50]  eta: 0:00:54  loss: 2.1775 (2.1679)  acc1: 54.4000 (53.4095)  acc5: 76.8000 (77.2571)  time: 1.2643  data: 1.2431  max mem: 2905
Test:  [30/50]  eta: 0:00:30  loss: 2.2345 (2.1959)  acc1: 51.2000 (52.7226)  acc5: 76.8000 (76.8516)  time: 1.1012  data: 1.0819  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 2.3541 (2.2453)  acc1: 50.4000 (51.9024)  acc5: 74.4000 (76.1951)  time: 0.6042  data: 0.5850  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.3377 (2.2435)  acc1: 51.2000 (51.9360)  acc5: 74.4000 (76.2400)  time: 0.6075  data: 0.5885  max mem: 2905
Test: Total time: 0:00:51 (1.0258 s / it)
* Acc@1 52.896 Acc@5 76.974 loss 2.217
Accuracy of the model on the 50000 test images: 52.9%
Max accuracy: 55.47%
Epoch: [220]  [  0/625]  eta: 3:49:14  lr: 0.000754  min_lr: 0.000754  loss: 2.8331 (2.8331)  class_acc: 0.5469 (0.5469)  weight_decay: 0.0500 (0.0500)  time: 22.0069  data: 19.7209  max mem: 2905
Epoch: [220]  [200/625]  eta: 0:14:19  lr: 0.000748  min_lr: 0.000748  loss: 2.7866 (2.8227)  class_acc: 0.5781 (0.5676)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5551 (3.4210)  time: 2.0994  data: 0.0007  max mem: 2905
Epoch: [220]  [400/625]  eta: 0:07:23  lr: 0.000743  min_lr: 0.000743  loss: 2.8310 (2.8385)  class_acc: 0.5625 (0.5629)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5217 (inf)  time: 1.9486  data: 0.0495  max mem: 2905
Epoch: [220]  [600/625]  eta: 0:00:49  lr: 0.000737  min_lr: 0.000737  loss: 2.9002 (2.8402)  class_acc: 0.5547 (0.5627)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.1108 (inf)  time: 2.0586  data: 0.0007  max mem: 2905
Epoch: [220]  [624/625]  eta: 0:00:01  lr: 0.000736  min_lr: 0.000736  loss: 2.8853 (2.8412)  class_acc: 0.5508 (0.5624)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.5105 (inf)  time: 0.4050  data: 0.0015  max mem: 2905
Epoch: [220] Total time: 0:20:11 (1.9384 s / it)
Averaged stats: lr: 0.000736  min_lr: 0.000736  loss: 2.8853 (2.8362)  class_acc: 0.5508 (0.5635)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.5105 (inf)
Test:  [ 0/50]  eta: 0:10:08  loss: 2.5442 (2.5442)  acc1: 48.8000 (48.8000)  acc5: 73.6000 (73.6000)  time: 12.1652  data: 12.1342  max mem: 2905
Test:  [10/50]  eta: 0:01:20  loss: 2.3319 (2.3187)  acc1: 51.2000 (52.2182)  acc5: 74.4000 (74.7636)  time: 2.0241  data: 2.0037  max mem: 2905
Test:  [20/50]  eta: 0:00:46  loss: 2.4589 (2.4447)  acc1: 49.6000 (49.0667)  acc5: 72.8000 (73.6381)  time: 1.0061  data: 0.9861  max mem: 2905
Test:  [30/50]  eta: 0:00:25  loss: 2.5015 (2.4221)  acc1: 49.6000 (49.8839)  acc5: 73.6000 (74.1936)  time: 0.8924  data: 0.8722  max mem: 2905
Test:  [40/50]  eta: 0:00:10  loss: 2.4818 (2.4485)  acc1: 50.4000 (49.2683)  acc5: 72.8000 (73.2293)  time: 0.5313  data: 0.5116  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 2.5120 (2.4597)  acc1: 46.4000 (49.0880)  acc5: 71.2000 (73.2160)  time: 0.4708  data: 0.4508  max mem: 2905
Test: Total time: 0:00:44 (0.8869 s / it)
* Acc@1 49.028 Acc@5 73.226 loss 2.430
Accuracy of the model on the 50000 test images: 49.0%
Max accuracy: 55.47%
Epoch: [221]  [  0/625]  eta: 3:24:02  lr: 0.000736  min_lr: 0.000736  loss: 2.7191 (2.7191)  class_acc: 0.6094 (0.6094)  weight_decay: 0.0500 (0.0500)  time: 19.5875  data: 19.2361  max mem: 2905
Epoch: [221]  [200/625]  eta: 0:14:19  lr: 0.000731  min_lr: 0.000731  loss: 2.7883 (2.8130)  class_acc: 0.5742 (0.5692)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.8464 (3.3130)  time: 1.8483  data: 0.0009  max mem: 2905
Epoch: [221]  [400/625]  eta: 0:07:25  lr: 0.000725  min_lr: 0.000725  loss: 2.8561 (2.8254)  class_acc: 0.5625 (0.5677)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7967 (3.2403)  time: 1.8415  data: 0.0009  max mem: 2905
Epoch: [221]  [600/625]  eta: 0:00:49  lr: 0.000720  min_lr: 0.000720  loss: 2.8826 (2.8304)  class_acc: 0.5547 (0.5670)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7030 (3.1994)  time: 2.0911  data: 0.0007  max mem: 2905
Epoch: [221]  [624/625]  eta: 0:00:01  lr: 0.000719  min_lr: 0.000719  loss: 2.8839 (2.8314)  class_acc: 0.5508 (0.5664)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3692 (3.1862)  time: 0.8245  data: 0.0014  max mem: 2905
Epoch: [221] Total time: 0:20:18 (1.9495 s / it)
Averaged stats: lr: 0.000719  min_lr: 0.000719  loss: 2.8839 (2.8317)  class_acc: 0.5508 (0.5652)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3692 (3.1862)
Test:  [ 0/50]  eta: 0:11:05  loss: 2.1198 (2.1198)  acc1: 48.8000 (48.8000)  acc5: 80.0000 (80.0000)  time: 13.3195  data: 13.2905  max mem: 2905
Test:  [10/50]  eta: 0:01:31  loss: 2.1198 (2.1117)  acc1: 53.6000 (53.4545)  acc5: 76.8000 (76.0000)  time: 2.2865  data: 2.2674  max mem: 2905
Test:  [20/50]  eta: 0:00:55  loss: 2.2780 (2.2737)  acc1: 50.4000 (50.5905)  acc5: 74.4000 (75.0857)  time: 1.2596  data: 1.2406  max mem: 2905
Test:  [30/50]  eta: 0:00:32  loss: 2.3160 (2.2701)  acc1: 50.4000 (51.6645)  acc5: 74.4000 (75.0710)  time: 1.2884  data: 1.2691  max mem: 2905
Test:  [40/50]  eta: 0:00:13  loss: 2.2639 (2.2961)  acc1: 52.0000 (51.2000)  acc5: 73.6000 (74.8098)  time: 0.9322  data: 0.9101  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.2748 (2.2995)  acc1: 47.2000 (51.0880)  acc5: 74.4000 (74.8960)  time: 0.8987  data: 0.8758  max mem: 2905
Test: Total time: 0:00:58 (1.1621 s / it)
* Acc@1 51.604 Acc@5 75.564 loss 2.276
Accuracy of the model on the 50000 test images: 51.6%
Max accuracy: 55.47%
Epoch: [222]  [  0/625]  eta: 3:33:08  lr: 0.000719  min_lr: 0.000719  loss: 2.7429 (2.7429)  class_acc: 0.5742 (0.5742)  weight_decay: 0.0500 (0.0500)  time: 20.4617  data: 17.8072  max mem: 2905
Epoch: [222]  [200/625]  eta: 0:14:27  lr: 0.000714  min_lr: 0.000714  loss: 2.8334 (2.8090)  class_acc: 0.5664 (0.5703)  weight_decay: 0.0500 (0.0500)  grad_norm: 4.6817 (3.6760)  time: 2.0688  data: 0.0077  max mem: 2905
Epoch: [222]  [400/625]  eta: 0:07:22  lr: 0.000708  min_lr: 0.000708  loss: 2.8283 (2.8220)  class_acc: 0.5586 (0.5670)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.3044 (3.7010)  time: 1.9727  data: 0.0009  max mem: 2905
Epoch: [222]  [600/625]  eta: 0:00:49  lr: 0.000703  min_lr: 0.000703  loss: 2.7872 (2.8256)  class_acc: 0.5625 (0.5660)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5118 (3.6341)  time: 1.9153  data: 0.0006  max mem: 2905
Epoch: [222]  [624/625]  eta: 0:00:01  lr: 0.000702  min_lr: 0.000702  loss: 2.8644 (2.8263)  class_acc: 0.5625 (0.5660)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5995 (3.5931)  time: 0.8339  data: 0.0014  max mem: 2905
Epoch: [222] Total time: 0:20:05 (1.9281 s / it)
Averaged stats: lr: 0.000702  min_lr: 0.000702  loss: 2.8644 (2.8276)  class_acc: 0.5625 (0.5657)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5995 (3.5931)
Test:  [ 0/50]  eta: 0:10:24  loss: 2.0144 (2.0144)  acc1: 52.8000 (52.8000)  acc5: 79.2000 (79.2000)  time: 12.4907  data: 12.4628  max mem: 2905
Test:  [10/50]  eta: 0:01:21  loss: 1.7840 (1.7650)  acc1: 59.2000 (59.6364)  acc5: 81.6000 (82.4727)  time: 2.0255  data: 2.0065  max mem: 2905
Test:  [20/50]  eta: 0:00:47  loss: 1.8505 (1.9359)  acc1: 56.0000 (56.8381)  acc5: 81.6000 (80.4191)  time: 1.0336  data: 1.0148  max mem: 2905
Test:  [30/50]  eta: 0:00:26  loss: 2.0733 (1.9769)  acc1: 53.6000 (56.3097)  acc5: 78.4000 (79.6903)  time: 0.9340  data: 0.9128  max mem: 2905
Test:  [40/50]  eta: 0:00:10  loss: 2.0218 (1.9949)  acc1: 54.4000 (56.0585)  acc5: 78.4000 (79.2195)  time: 0.5812  data: 0.5604  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 2.0308 (2.0174)  acc1: 55.2000 (55.6160)  acc5: 77.6000 (78.9440)  time: 0.5745  data: 0.5565  max mem: 2905
Test: Total time: 0:00:46 (0.9377 s / it)
* Acc@1 56.632 Acc@5 79.536 loss 1.973
Accuracy of the model on the 50000 test images: 56.6%
Max accuracy: 56.63%
Epoch: [223]  [  0/625]  eta: 3:22:34  lr: 0.000702  min_lr: 0.000702  loss: 2.9567 (2.9567)  class_acc: 0.5391 (0.5391)  weight_decay: 0.0500 (0.0500)  time: 19.4466  data: 19.2668  max mem: 2905
Epoch: [223]  [200/625]  eta: 0:13:51  lr: 0.000696  min_lr: 0.000696  loss: 2.8644 (2.8272)  class_acc: 0.5664 (0.5657)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.6796 (3.6625)  time: 1.7534  data: 0.2931  max mem: 2905
Epoch: [223]  [400/625]  eta: 0:07:07  lr: 0.000691  min_lr: 0.000691  loss: 2.8362 (2.8247)  class_acc: 0.5547 (0.5659)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.2313 (3.3947)  time: 1.8213  data: 0.0668  max mem: 2905
Epoch: [223]  [600/625]  eta: 0:00:47  lr: 0.000686  min_lr: 0.000686  loss: 2.8040 (2.8253)  class_acc: 0.5625 (0.5663)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.9826 (3.3668)  time: 1.7651  data: 0.3873  max mem: 2905
Epoch: [223]  [624/625]  eta: 0:00:01  lr: 0.000685  min_lr: 0.000685  loss: 2.8220 (2.8259)  class_acc: 0.5625 (0.5662)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3721 (3.3306)  time: 0.7192  data: 0.0981  max mem: 2905
Epoch: [223] Total time: 0:19:33 (1.8772 s / it)
Averaged stats: lr: 0.000685  min_lr: 0.000685  loss: 2.8220 (2.8278)  class_acc: 0.5625 (0.5659)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3721 (3.3306)
Test:  [ 0/50]  eta: 0:09:39  loss: 2.2917 (2.2917)  acc1: 54.4000 (54.4000)  acc5: 72.0000 (72.0000)  time: 11.5949  data: 11.5598  max mem: 2905
Test:  [10/50]  eta: 0:01:19  loss: 2.1529 (2.0980)  acc1: 53.6000 (54.4727)  acc5: 76.8000 (76.8000)  time: 1.9888  data: 1.9659  max mem: 2905
Test:  [20/50]  eta: 0:00:47  loss: 2.1606 (2.1968)  acc1: 51.2000 (52.8000)  acc5: 76.8000 (76.6857)  time: 1.0899  data: 1.0689  max mem: 2905
Test:  [30/50]  eta: 0:00:27  loss: 2.3104 (2.2169)  acc1: 51.2000 (52.5419)  acc5: 76.8000 (76.1806)  time: 1.0420  data: 1.0200  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 2.3458 (2.2673)  acc1: 50.4000 (51.8439)  acc5: 75.2000 (75.3561)  time: 0.6473  data: 0.6240  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 2.3437 (2.2792)  acc1: 50.4000 (51.7920)  acc5: 73.6000 (75.1040)  time: 0.6055  data: 0.5842  max mem: 2905
Test: Total time: 0:00:46 (0.9381 s / it)
* Acc@1 52.020 Acc@5 76.140 loss 2.233
Accuracy of the model on the 50000 test images: 52.0%
Max accuracy: 56.63%
Epoch: [224]  [  0/625]  eta: 3:18:55  lr: 0.000685  min_lr: 0.000685  loss: 2.7990 (2.7990)  class_acc: 0.5898 (0.5898)  weight_decay: 0.0500 (0.0500)  time: 19.0963  data: 18.9780  max mem: 2905
Epoch: [224]  [200/625]  eta: 0:13:42  lr: 0.000680  min_lr: 0.000680  loss: 2.7993 (2.8212)  class_acc: 0.5625 (0.5712)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4578 (3.1635)  time: 1.9367  data: 0.0007  max mem: 2905
Epoch: [224]  [400/625]  eta: 0:07:12  lr: 0.000674  min_lr: 0.000674  loss: 2.8184 (2.8224)  class_acc: 0.5664 (0.5691)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6766 (3.3197)  time: 1.9716  data: 0.0663  max mem: 2905
Epoch: [224]  [600/625]  eta: 0:00:48  lr: 0.000669  min_lr: 0.000669  loss: 2.8155 (2.8274)  class_acc: 0.5664 (0.5676)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1413 (3.2610)  time: 1.8458  data: 0.0329  max mem: 2905
Epoch: [224]  [624/625]  eta: 0:00:01  lr: 0.000668  min_lr: 0.000668  loss: 2.8499 (2.8284)  class_acc: 0.5430 (0.5673)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.3804 (3.2899)  time: 0.8596  data: 0.0259  max mem: 2905
Epoch: [224] Total time: 0:19:38 (1.8864 s / it)
Averaged stats: lr: 0.000668  min_lr: 0.000668  loss: 2.8499 (2.8211)  class_acc: 0.5430 (0.5675)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.3804 (3.2899)
Test:  [ 0/50]  eta: 0:09:54  loss: 2.1635 (2.1635)  acc1: 52.8000 (52.8000)  acc5: 80.0000 (80.0000)  time: 11.8913  data: 11.8653  max mem: 2905
Test:  [10/50]  eta: 0:01:23  loss: 1.9494 (1.9558)  acc1: 56.0000 (56.5091)  acc5: 80.8000 (80.5818)  time: 2.0800  data: 2.0612  max mem: 2905
Test:  [20/50]  eta: 0:00:49  loss: 2.0537 (2.1261)  acc1: 51.2000 (53.2191)  acc5: 77.6000 (78.5524)  time: 1.1533  data: 1.1341  max mem: 2905
Test:  [30/50]  eta: 0:00:30  loss: 2.3219 (2.1630)  acc1: 51.2000 (53.0323)  acc5: 76.0000 (77.7290)  time: 1.1852  data: 1.1658  max mem: 2905
Test:  [40/50]  eta: 0:00:13  loss: 2.1732 (2.1829)  acc1: 51.2000 (52.8195)  acc5: 76.0000 (77.2293)  time: 0.9708  data: 0.9524  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.1732 (2.1901)  acc1: 50.4000 (52.4960)  acc5: 76.0000 (76.9920)  time: 0.8942  data: 0.8750  max mem: 2905
Test: Total time: 0:00:55 (1.1077 s / it)
* Acc@1 53.376 Acc@5 77.586 loss 2.148
Accuracy of the model on the 50000 test images: 53.4%
Max accuracy: 56.63%
Epoch: [225]  [  0/625]  eta: 3:50:12  lr: 0.000668  min_lr: 0.000668  loss: 2.7869 (2.7869)  class_acc: 0.5391 (0.5391)  weight_decay: 0.0500 (0.0500)  time: 22.1005  data: 16.8768  max mem: 2905
Epoch: [225]  [200/625]  eta: 0:13:42  lr: 0.000663  min_lr: 0.000663  loss: 2.8162 (2.8060)  class_acc: 0.5625 (0.5691)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5242 (2.9538)  time: 1.6414  data: 0.1694  max mem: 2905
Epoch: [225]  [400/625]  eta: 0:07:11  lr: 0.000657  min_lr: 0.000657  loss: 2.8392 (2.8119)  class_acc: 0.5586 (0.5688)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.2505 (3.1525)  time: 1.9047  data: 0.0820  max mem: 2905
Epoch: [225]  [600/625]  eta: 0:00:47  lr: 0.000652  min_lr: 0.000652  loss: 2.8350 (2.8188)  class_acc: 0.5625 (0.5675)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4554 (3.0685)  time: 1.8824  data: 0.8664  max mem: 2905
Epoch: [225]  [624/625]  eta: 0:00:01  lr: 0.000652  min_lr: 0.000652  loss: 2.8253 (2.8200)  class_acc: 0.5742 (0.5674)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1150 (3.0519)  time: 0.7363  data: 0.2715  max mem: 2905
Epoch: [225] Total time: 0:19:44 (1.8953 s / it)
Averaged stats: lr: 0.000652  min_lr: 0.000652  loss: 2.8253 (2.8192)  class_acc: 0.5742 (0.5679)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1150 (3.0519)
Test:  [ 0/50]  eta: 0:09:34  loss: 1.6075 (1.6075)  acc1: 57.6000 (57.6000)  acc5: 88.8000 (88.8000)  time: 11.4850  data: 11.4609  max mem: 2905
Test:  [10/50]  eta: 0:01:19  loss: 1.7576 (1.7273)  acc1: 60.0000 (60.3636)  acc5: 84.0000 (83.7818)  time: 1.9831  data: 1.9645  max mem: 2905
Test:  [20/50]  eta: 0:00:47  loss: 1.8631 (1.8731)  acc1: 59.2000 (58.2857)  acc5: 80.8000 (81.6000)  time: 1.0916  data: 1.0735  max mem: 2905
Test:  [30/50]  eta: 0:00:27  loss: 1.9401 (1.8816)  acc1: 56.0000 (58.1677)  acc5: 80.8000 (81.6000)  time: 1.0413  data: 1.0228  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 1.9401 (1.8972)  acc1: 56.0000 (57.9122)  acc5: 81.6000 (81.1317)  time: 0.6668  data: 0.6452  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 1.9292 (1.9028)  acc1: 55.2000 (57.6160)  acc5: 80.8000 (81.0560)  time: 0.6076  data: 0.5857  max mem: 2905
Test: Total time: 0:00:47 (0.9564 s / it)
* Acc@1 58.326 Acc@5 81.262 loss 1.876
Accuracy of the model on the 50000 test images: 58.3%
Max accuracy: 58.33%
Epoch: [226]  [  0/625]  eta: 3:19:33  lr: 0.000651  min_lr: 0.000651  loss: 2.7994 (2.7994)  class_acc: 0.5820 (0.5820)  weight_decay: 0.0500 (0.0500)  time: 19.1581  data: 18.9488  max mem: 2905
Epoch: [226]  [200/625]  eta: 0:13:54  lr: 0.000646  min_lr: 0.000646  loss: 2.7681 (2.7975)  class_acc: 0.5742 (0.5737)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.0465 (3.2634)  time: 1.9344  data: 0.0008  max mem: 2905
Epoch: [226]  [400/625]  eta: 0:07:16  lr: 0.000641  min_lr: 0.000641  loss: 2.7806 (2.8149)  class_acc: 0.5742 (0.5704)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.1574 (3.2118)  time: 1.9394  data: 0.0457  max mem: 2905
Epoch: [226]  [600/625]  eta: 0:00:48  lr: 0.000636  min_lr: 0.000636  loss: 2.8242 (2.8189)  class_acc: 0.5586 (0.5685)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.6197 (3.4137)  time: 1.7290  data: 0.0008  max mem: 2905
Epoch: [226]  [624/625]  eta: 0:00:01  lr: 0.000635  min_lr: 0.000635  loss: 2.8306 (2.8193)  class_acc: 0.5625 (0.5685)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4763 (3.3777)  time: 0.8559  data: 0.0014  max mem: 2905
Epoch: [226] Total time: 0:19:46 (1.8983 s / it)
Averaged stats: lr: 0.000635  min_lr: 0.000635  loss: 2.8306 (2.8159)  class_acc: 0.5625 (0.5684)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4763 (3.3777)
Test:  [ 0/50]  eta: 0:10:12  loss: 2.1853 (2.1853)  acc1: 49.6000 (49.6000)  acc5: 77.6000 (77.6000)  time: 12.2435  data: 12.2151  max mem: 2905
Test:  [10/50]  eta: 0:01:21  loss: 2.0433 (2.0694)  acc1: 53.6000 (54.6182)  acc5: 77.6000 (77.4545)  time: 2.0473  data: 2.0272  max mem: 2905
Test:  [20/50]  eta: 0:00:47  loss: 2.2561 (2.2374)  acc1: 51.2000 (52.0000)  acc5: 75.2000 (75.6571)  time: 1.0616  data: 1.0413  max mem: 2905
Test:  [30/50]  eta: 0:00:28  loss: 2.3767 (2.2360)  acc1: 50.4000 (52.3613)  acc5: 75.2000 (75.8194)  time: 1.0745  data: 1.0535  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 2.2327 (2.2545)  acc1: 51.2000 (51.6488)  acc5: 75.2000 (75.3951)  time: 0.8368  data: 0.8164  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.2977 (2.2851)  acc1: 47.2000 (50.8000)  acc5: 74.4000 (75.0240)  time: 0.7598  data: 0.7395  max mem: 2905
Test: Total time: 0:00:51 (1.0249 s / it)
* Acc@1 51.768 Acc@5 75.880 loss 2.233
Accuracy of the model on the 50000 test images: 51.8%
Max accuracy: 58.33%
Epoch: [227]  [  0/625]  eta: 4:23:18  lr: 0.000635  min_lr: 0.000635  loss: 2.7508 (2.7508)  class_acc: 0.6289 (0.6289)  weight_decay: 0.0500 (0.0500)  time: 25.2776  data: 21.0737  max mem: 2905
Epoch: [227]  [200/625]  eta: 0:14:30  lr: 0.000630  min_lr: 0.000630  loss: 2.8346 (2.8022)  class_acc: 0.5586 (0.5686)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.8335 (3.4276)  time: 2.0562  data: 0.0695  max mem: 2905
Epoch: [227]  [400/625]  eta: 0:07:23  lr: 0.000625  min_lr: 0.000625  loss: 2.8212 (2.8047)  class_acc: 0.5664 (0.5695)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7345 (3.3682)  time: 1.8246  data: 0.0554  max mem: 2905
Epoch: [227]  [600/625]  eta: 0:00:49  lr: 0.000619  min_lr: 0.000619  loss: 2.7757 (2.8087)  class_acc: 0.5742 (0.5687)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4021 (3.4449)  time: 2.0238  data: 0.0033  max mem: 2905
Epoch: [227]  [624/625]  eta: 0:00:01  lr: 0.000619  min_lr: 0.000619  loss: 2.7762 (2.8077)  class_acc: 0.5703 (0.5690)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3825 (3.4170)  time: 0.7313  data: 0.0013  max mem: 2905
Epoch: [227] Total time: 0:19:56 (1.9137 s / it)
Averaged stats: lr: 0.000619  min_lr: 0.000619  loss: 2.7762 (2.8117)  class_acc: 0.5703 (0.5693)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3825 (3.4170)
Test:  [ 0/50]  eta: 0:09:40  loss: 1.9122 (1.9122)  acc1: 58.4000 (58.4000)  acc5: 82.4000 (82.4000)  time: 11.6133  data: 11.5741  max mem: 2905
Test:  [10/50]  eta: 0:01:12  loss: 1.8697 (1.7995)  acc1: 58.4000 (59.2727)  acc5: 80.8000 (81.3818)  time: 1.8095  data: 1.7879  max mem: 2905
Test:  [20/50]  eta: 0:00:41  loss: 1.9306 (1.9705)  acc1: 56.0000 (56.4191)  acc5: 78.4000 (79.7714)  time: 0.8581  data: 0.8385  max mem: 2905
Test:  [30/50]  eta: 0:00:26  loss: 2.1099 (2.0106)  acc1: 52.8000 (55.5355)  acc5: 78.4000 (79.2774)  time: 1.0661  data: 1.0466  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 2.0536 (2.0259)  acc1: 52.8000 (55.3561)  acc5: 78.4000 (78.8878)  time: 0.9853  data: 0.9665  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.0485 (2.0400)  acc1: 54.4000 (54.9600)  acc5: 78.4000 (78.8160)  time: 0.5242  data: 0.5047  max mem: 2905
Test: Total time: 0:00:51 (1.0377 s / it)
* Acc@1 55.902 Acc@5 79.566 loss 1.993
Accuracy of the model on the 50000 test images: 55.9%
Max accuracy: 58.33%
Epoch: [228]  [  0/625]  eta: 3:33:30  lr: 0.000619  min_lr: 0.000619  loss: 2.8542 (2.8542)  class_acc: 0.5312 (0.5312)  weight_decay: 0.0500 (0.0500)  time: 20.4963  data: 19.0074  max mem: 2905
Epoch: [228]  [200/625]  eta: 0:14:04  lr: 0.000614  min_lr: 0.000614  loss: 2.8009 (2.8050)  class_acc: 0.5781 (0.5716)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.0872 (3.1895)  time: 1.8658  data: 1.6852  max mem: 2905
Epoch: [228]  [400/625]  eta: 0:07:20  lr: 0.000608  min_lr: 0.000608  loss: 2.7718 (2.8085)  class_acc: 0.5742 (0.5702)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.0361 (3.2611)  time: 1.8774  data: 0.0008  max mem: 2905
Epoch: [228]  [600/625]  eta: 0:00:49  lr: 0.000603  min_lr: 0.000603  loss: 2.7492 (2.8101)  class_acc: 0.5820 (0.5692)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.3052 (3.2620)  time: 1.8297  data: 0.0008  max mem: 2905
Epoch: [228]  [624/625]  eta: 0:00:01  lr: 0.000603  min_lr: 0.000603  loss: 2.8008 (2.8095)  class_acc: 0.5703 (0.5694)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.4762 (3.2899)  time: 0.8699  data: 0.0187  max mem: 2905
Epoch: [228] Total time: 0:20:07 (1.9318 s / it)
Averaged stats: lr: 0.000603  min_lr: 0.000603  loss: 2.8008 (2.8107)  class_acc: 0.5703 (0.5699)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.4762 (3.2899)
Test:  [ 0/50]  eta: 0:10:58  loss: 1.8527 (1.8527)  acc1: 57.6000 (57.6000)  acc5: 83.2000 (83.2000)  time: 13.1633  data: 13.1382  max mem: 2905
Test:  [10/50]  eta: 0:01:33  loss: 1.7647 (1.7371)  acc1: 60.0000 (60.5091)  acc5: 81.6000 (82.6909)  time: 2.3490  data: 2.3300  max mem: 2905
Test:  [20/50]  eta: 0:00:56  loss: 1.9968 (1.9506)  acc1: 56.0000 (56.6476)  acc5: 79.2000 (80.8000)  time: 1.3055  data: 1.2864  max mem: 2905
Test:  [30/50]  eta: 0:00:32  loss: 2.1427 (1.9518)  acc1: 55.2000 (57.1871)  acc5: 79.2000 (80.5936)  time: 1.2016  data: 1.1829  max mem: 2905
Test:  [40/50]  eta: 0:00:13  loss: 1.8928 (1.9603)  acc1: 56.0000 (57.2878)  acc5: 80.8000 (80.2342)  time: 0.7832  data: 0.7653  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 1.9211 (1.9747)  acc1: 56.0000 (56.7840)  acc5: 79.2000 (79.9680)  time: 0.6976  data: 0.6799  max mem: 2905
Test: Total time: 0:00:56 (1.1202 s / it)
* Acc@1 57.990 Acc@5 80.910 loss 1.928
Accuracy of the model on the 50000 test images: 58.0%
Max accuracy: 58.33%
Epoch: [229]  [  0/625]  eta: 4:17:54  lr: 0.000603  min_lr: 0.000603  loss: 2.6692 (2.6692)  class_acc: 0.5898 (0.5898)  weight_decay: 0.0500 (0.0500)  time: 24.7595  data: 19.2615  max mem: 2905
Epoch: [229]  [200/625]  eta: 0:14:13  lr: 0.000597  min_lr: 0.000597  loss: 2.7499 (2.7962)  class_acc: 0.5703 (0.5730)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7308 (3.5185)  time: 1.8759  data: 0.0009  max mem: 2905
Epoch: [229]  [400/625]  eta: 0:07:20  lr: 0.000592  min_lr: 0.000592  loss: 2.8252 (2.7999)  class_acc: 0.5703 (0.5718)  weight_decay: 0.0500 (0.0500)  grad_norm: 4.5798 (3.5899)  time: 1.9614  data: 0.0145  max mem: 2905
Epoch: [229]  [600/625]  eta: 0:00:49  lr: 0.000587  min_lr: 0.000587  loss: 2.8627 (2.8065)  class_acc: 0.5547 (0.5707)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.3104 (3.6582)  time: 2.0219  data: 0.0376  max mem: 2905
Epoch: [229]  [624/625]  eta: 0:00:01  lr: 0.000587  min_lr: 0.000587  loss: 2.8191 (2.8063)  class_acc: 0.5742 (0.5709)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.8051 (3.6175)  time: 0.7761  data: 0.0014  max mem: 2905
Epoch: [229] Total time: 0:20:15 (1.9449 s / it)
Averaged stats: lr: 0.000587  min_lr: 0.000587  loss: 2.8191 (2.8083)  class_acc: 0.5742 (0.5704)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.8051 (3.6175)
Test:  [ 0/50]  eta: 0:11:08  loss: 2.1459 (2.1459)  acc1: 52.8000 (52.8000)  acc5: 77.6000 (77.6000)  time: 13.3742  data: 13.3357  max mem: 2905
Test:  [10/50]  eta: 0:01:30  loss: 1.9254 (1.9042)  acc1: 59.2000 (59.4182)  acc5: 80.8000 (81.0182)  time: 2.2651  data: 2.2440  max mem: 2905
Test:  [20/50]  eta: 0:00:53  loss: 1.9960 (2.0638)  acc1: 56.8000 (55.1619)  acc5: 79.2000 (78.9333)  time: 1.2070  data: 1.1861  max mem: 2905
Test:  [30/50]  eta: 0:00:31  loss: 2.1468 (2.0659)  acc1: 52.0000 (54.9161)  acc5: 77.6000 (78.7613)  time: 1.2320  data: 1.2117  max mem: 2905
Test:  [40/50]  eta: 0:00:13  loss: 2.1743 (2.1175)  acc1: 52.8000 (53.8341)  acc5: 76.8000 (78.0878)  time: 0.8127  data: 0.7917  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.2134 (2.1298)  acc1: 52.8000 (53.7920)  acc5: 76.8000 (78.0000)  time: 0.7434  data: 0.7218  max mem: 2905
Test: Total time: 0:00:55 (1.1021 s / it)
* Acc@1 54.580 Acc@5 78.262 loss 2.089
Accuracy of the model on the 50000 test images: 54.6%
Max accuracy: 58.33%
Epoch: [230]  [  0/625]  eta: 4:16:09  lr: 0.000587  min_lr: 0.000587  loss: 3.1090 (3.1090)  class_acc: 0.4883 (0.4883)  weight_decay: 0.0500 (0.0500)  time: 24.5911  data: 18.7988  max mem: 2905
Epoch: [230]  [200/625]  eta: 0:14:42  lr: 0.000582  min_lr: 0.000582  loss: 2.7681 (2.7908)  class_acc: 0.5820 (0.5754)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.9388 (3.5806)  time: 2.0226  data: 0.0008  max mem: 2905
Epoch: [230]  [400/625]  eta: 0:07:32  lr: 0.000577  min_lr: 0.000577  loss: 2.7967 (2.8021)  class_acc: 0.5742 (0.5725)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.0681 (3.2917)  time: 2.0449  data: 0.0006  max mem: 2905
Epoch: [230]  [600/625]  eta: 0:00:50  lr: 0.000571  min_lr: 0.000571  loss: 2.8194 (2.8068)  class_acc: 0.5703 (0.5708)  weight_decay: 0.0500 (0.0500)  grad_norm: 4.1883 (3.3142)  time: 2.1214  data: 0.0008  max mem: 2905
Epoch: [230]  [624/625]  eta: 0:00:01  lr: 0.000571  min_lr: 0.000571  loss: 2.8307 (2.8076)  class_acc: 0.5664 (0.5707)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.9054 (3.3572)  time: 0.7775  data: 0.0017  max mem: 2905
Epoch: [230] Total time: 0:20:21 (1.9552 s / it)
Averaged stats: lr: 0.000571  min_lr: 0.000571  loss: 2.8307 (2.8048)  class_acc: 0.5664 (0.5709)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.9054 (3.3572)
Test:  [ 0/50]  eta: 0:10:41  loss: 1.8700 (1.8700)  acc1: 55.2000 (55.2000)  acc5: 82.4000 (82.4000)  time: 12.8291  data: 12.8047  max mem: 2905
Test:  [10/50]  eta: 0:01:23  loss: 1.8156 (1.8103)  acc1: 61.6000 (60.8000)  acc5: 82.4000 (82.2545)  time: 2.0979  data: 2.0777  max mem: 2905
Test:  [20/50]  eta: 0:00:48  loss: 2.0411 (1.9976)  acc1: 57.6000 (56.3048)  acc5: 78.4000 (80.0000)  time: 1.0680  data: 1.0481  max mem: 2905
Test:  [30/50]  eta: 0:00:28  loss: 2.1375 (2.0355)  acc1: 52.8000 (55.9226)  acc5: 77.6000 (78.9677)  time: 1.0494  data: 1.0295  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 2.0447 (2.0545)  acc1: 53.6000 (55.9024)  acc5: 76.0000 (78.4000)  time: 0.6936  data: 0.6748  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 2.0692 (2.0548)  acc1: 55.2000 (55.9360)  acc5: 76.8000 (78.4320)  time: 0.6029  data: 0.5848  max mem: 2905
Test: Total time: 0:00:49 (0.9855 s / it)
* Acc@1 56.292 Acc@5 79.016 loss 2.032
Accuracy of the model on the 50000 test images: 56.3%
Max accuracy: 58.33%
Epoch: [231]  [  0/625]  eta: 3:19:52  lr: 0.000571  min_lr: 0.000571  loss: 2.8739 (2.8739)  class_acc: 0.5430 (0.5430)  weight_decay: 0.0500 (0.0500)  time: 19.1873  data: 17.8375  max mem: 2905
Epoch: [231]  [200/625]  eta: 0:14:09  lr: 0.000566  min_lr: 0.000566  loss: 2.7872 (2.7965)  class_acc: 0.5703 (0.5742)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.9912 (3.2443)  time: 1.9974  data: 0.9620  max mem: 2905
Epoch: [231]  [400/625]  eta: 0:07:15  lr: 0.000561  min_lr: 0.000561  loss: 2.8260 (2.8033)  class_acc: 0.5664 (0.5726)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.0155 (inf)  time: 1.8714  data: 1.6883  max mem: 2905
Epoch: [231]  [600/625]  eta: 0:00:48  lr: 0.000556  min_lr: 0.000556  loss: 2.7760 (2.8042)  class_acc: 0.5664 (0.5721)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7760 (inf)  time: 1.7572  data: 1.5709  max mem: 2905
Epoch: [231]  [624/625]  eta: 0:00:01  lr: 0.000555  min_lr: 0.000555  loss: 2.7955 (2.8046)  class_acc: 0.5703 (0.5721)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2778 (inf)  time: 0.6970  data: 0.5400  max mem: 2905
Epoch: [231] Total time: 0:19:46 (1.8979 s / it)
Averaged stats: lr: 0.000555  min_lr: 0.000555  loss: 2.7955 (2.8014)  class_acc: 0.5703 (0.5715)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2778 (inf)
Test:  [ 0/50]  eta: 0:10:16  loss: 1.6774 (1.6774)  acc1: 63.2000 (63.2000)  acc5: 85.6000 (85.6000)  time: 12.3347  data: 12.3111  max mem: 2905
Test:  [10/50]  eta: 0:01:21  loss: 1.7365 (1.7937)  acc1: 61.6000 (60.4364)  acc5: 83.2000 (82.9091)  time: 2.0328  data: 2.0111  max mem: 2905
Test:  [20/50]  eta: 0:00:46  loss: 1.9281 (1.9538)  acc1: 55.2000 (57.0667)  acc5: 80.8000 (81.0286)  time: 1.0184  data: 0.9977  max mem: 2905
Test:  [30/50]  eta: 0:00:25  loss: 2.0676 (1.9567)  acc1: 54.4000 (57.1097)  acc5: 80.0000 (80.6452)  time: 0.8577  data: 0.8387  max mem: 2905
Test:  [40/50]  eta: 0:00:10  loss: 1.9374 (1.9935)  acc1: 55.2000 (56.5463)  acc5: 78.4000 (79.5902)  time: 0.4824  data: 0.4633  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 2.1219 (2.0049)  acc1: 54.4000 (56.5920)  acc5: 76.8000 (79.2640)  time: 0.4467  data: 0.4272  max mem: 2905
Test: Total time: 0:00:42 (0.8586 s / it)
* Acc@1 56.626 Acc@5 79.948 loss 1.972
Accuracy of the model on the 50000 test images: 56.6%
Max accuracy: 58.33%
Epoch: [232]  [  0/625]  eta: 4:07:03  lr: 0.000555  min_lr: 0.000555  loss: 2.8731 (2.8731)  class_acc: 0.5703 (0.5703)  weight_decay: 0.0500 (0.0500)  time: 23.7178  data: 17.3007  max mem: 2905
Epoch: [232]  [200/625]  eta: 0:13:42  lr: 0.000550  min_lr: 0.000550  loss: 2.7129 (2.7817)  class_acc: 0.5781 (0.5741)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.9729 (3.1949)  time: 1.4717  data: 0.0010  max mem: 2905
Epoch: [232]  [400/625]  eta: 0:07:13  lr: 0.000545  min_lr: 0.000545  loss: 2.7285 (2.7850)  class_acc: 0.5820 (0.5737)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.5392 (3.4313)  time: 2.1543  data: 0.0008  max mem: 2905
Epoch: [232]  [600/625]  eta: 0:00:48  lr: 0.000540  min_lr: 0.000540  loss: 2.7790 (2.7916)  class_acc: 0.5742 (0.5729)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.4367 (3.3541)  time: 1.9816  data: 0.0152  max mem: 2905
Epoch: [232]  [624/625]  eta: 0:00:01  lr: 0.000540  min_lr: 0.000540  loss: 2.7984 (2.7925)  class_acc: 0.5625 (0.5727)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3837 (3.3191)  time: 0.7159  data: 0.0018  max mem: 2905
Epoch: [232] Total time: 0:19:57 (1.9160 s / it)
Averaged stats: lr: 0.000540  min_lr: 0.000540  loss: 2.7984 (2.7976)  class_acc: 0.5625 (0.5725)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3837 (3.3191)
Test:  [ 0/50]  eta: 0:09:21  loss: 1.8810 (1.8810)  acc1: 56.8000 (56.8000)  acc5: 80.0000 (80.0000)  time: 11.2215  data: 11.1916  max mem: 2905
Test:  [10/50]  eta: 0:01:12  loss: 1.8810 (1.8578)  acc1: 59.2000 (59.2727)  acc5: 80.8000 (81.4545)  time: 1.8122  data: 1.7916  max mem: 2905
Test:  [20/50]  eta: 0:00:42  loss: 1.9956 (2.0134)  acc1: 54.4000 (56.0000)  acc5: 79.2000 (79.6191)  time: 0.9124  data: 0.8930  max mem: 2905
Test:  [30/50]  eta: 0:00:27  loss: 2.1664 (2.0434)  acc1: 52.8000 (55.8710)  acc5: 77.6000 (79.0194)  time: 1.1048  data: 1.0865  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 2.0152 (2.0610)  acc1: 54.4000 (55.2195)  acc5: 79.2000 (78.5951)  time: 1.0898  data: 1.0721  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 2.0864 (2.0574)  acc1: 54.4000 (55.4400)  acc5: 78.4000 (78.6240)  time: 0.6653  data: 0.6466  max mem: 2905
Test: Total time: 0:00:55 (1.1098 s / it)
* Acc@1 56.114 Acc@5 78.878 loss 2.028
Accuracy of the model on the 50000 test images: 56.1%
Max accuracy: 58.33%
Epoch: [233]  [  0/625]  eta: 4:17:12  lr: 0.000540  min_lr: 0.000540  loss: 2.6610 (2.6610)  class_acc: 0.5703 (0.5703)  weight_decay: 0.0500 (0.0500)  time: 24.6915  data: 18.3156  max mem: 2905
Epoch: [233]  [200/625]  eta: 0:14:28  lr: 0.000535  min_lr: 0.000535  loss: 2.7795 (2.7874)  class_acc: 0.5742 (0.5741)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.2357 (3.3048)  time: 1.8983  data: 0.0005  max mem: 2905
Epoch: [233]  [400/625]  eta: 0:07:18  lr: 0.000530  min_lr: 0.000530  loss: 2.7779 (2.7918)  class_acc: 0.5742 (0.5739)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.0097 (3.3251)  time: 1.8607  data: 1.4458  max mem: 2905
Epoch: [233]  [600/625]  eta: 0:00:48  lr: 0.000525  min_lr: 0.000525  loss: 2.7850 (2.7951)  class_acc: 0.5664 (0.5739)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.8634 (3.4273)  time: 2.0629  data: 0.0242  max mem: 2905
Epoch: [233]  [624/625]  eta: 0:00:01  lr: 0.000525  min_lr: 0.000525  loss: 2.8057 (2.7955)  class_acc: 0.5742 (0.5740)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6490 (3.3994)  time: 0.7271  data: 0.0012  max mem: 2905
Epoch: [233] Total time: 0:19:52 (1.9080 s / it)
Averaged stats: lr: 0.000525  min_lr: 0.000525  loss: 2.8057 (2.7961)  class_acc: 0.5742 (0.5733)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6490 (3.3994)
Test:  [ 0/50]  eta: 0:10:07  loss: 1.7354 (1.7354)  acc1: 62.4000 (62.4000)  acc5: 86.4000 (86.4000)  time: 12.1419  data: 12.1121  max mem: 2905
Test:  [10/50]  eta: 0:01:07  loss: 1.7092 (1.6220)  acc1: 64.0000 (64.2182)  acc5: 86.4000 (85.7455)  time: 1.6821  data: 1.6618  max mem: 2905
Test:  [20/50]  eta: 0:00:41  loss: 1.7951 (1.8257)  acc1: 57.6000 (59.4667)  acc5: 82.4000 (82.5143)  time: 0.8459  data: 0.8256  max mem: 2905
Test:  [30/50]  eta: 0:00:25  loss: 1.9529 (1.8486)  acc1: 55.2000 (59.1484)  acc5: 80.0000 (82.0129)  time: 1.0870  data: 1.0675  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 1.8771 (1.8581)  acc1: 56.8000 (59.0244)  acc5: 80.0000 (81.6000)  time: 0.8831  data: 0.8656  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 1.8965 (1.8725)  acc1: 56.8000 (58.8640)  acc5: 80.0000 (81.4880)  time: 0.4732  data: 0.4559  max mem: 2905
Test: Total time: 0:00:49 (0.9974 s / it)
* Acc@1 59.628 Acc@5 82.046 loss 1.835
Accuracy of the model on the 50000 test images: 59.6%
Max accuracy: 59.63%
Epoch: [234]  [  0/625]  eta: 4:11:41  lr: 0.000525  min_lr: 0.000525  loss: 2.7511 (2.7511)  class_acc: 0.6250 (0.6250)  weight_decay: 0.0500 (0.0500)  time: 24.1620  data: 19.2059  max mem: 2905
Epoch: [234]  [200/625]  eta: 0:14:28  lr: 0.000520  min_lr: 0.000520  loss: 2.7253 (2.7788)  class_acc: 0.5898 (0.5775)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6878 (3.1516)  time: 1.9309  data: 0.0015  max mem: 2905
Epoch: [234]  [400/625]  eta: 0:07:26  lr: 0.000515  min_lr: 0.000515  loss: 2.7892 (2.7906)  class_acc: 0.5703 (0.5753)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.2308 (3.3925)  time: 1.8564  data: 0.0011  max mem: 2905
Epoch: [234]  [600/625]  eta: 0:00:49  lr: 0.000510  min_lr: 0.000510  loss: 2.8073 (2.7955)  class_acc: 0.5703 (0.5738)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.2033 (3.2646)  time: 1.9608  data: 0.0011  max mem: 2905
Epoch: [234]  [624/625]  eta: 0:00:01  lr: 0.000510  min_lr: 0.000510  loss: 2.7745 (2.7959)  class_acc: 0.5703 (0.5737)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.8156 (3.2592)  time: 0.7573  data: 0.0014  max mem: 2905
Epoch: [234] Total time: 0:20:10 (1.9363 s / it)
Averaged stats: lr: 0.000510  min_lr: 0.000510  loss: 2.7745 (2.7935)  class_acc: 0.5703 (0.5739)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.8156 (3.2592)
Test:  [ 0/50]  eta: 0:09:36  loss: 1.7962 (1.7962)  acc1: 63.2000 (63.2000)  acc5: 83.2000 (83.2000)  time: 11.5244  data: 11.4893  max mem: 2905
Test:  [10/50]  eta: 0:01:18  loss: 1.6752 (1.7176)  acc1: 63.2000 (62.9091)  acc5: 83.2000 (84.0000)  time: 1.9734  data: 1.9518  max mem: 2905
Test:  [20/50]  eta: 0:00:47  loss: 1.8667 (1.8785)  acc1: 56.0000 (59.3524)  acc5: 81.6000 (82.0952)  time: 1.0771  data: 1.0573  max mem: 2905
Test:  [30/50]  eta: 0:00:27  loss: 1.8704 (1.8811)  acc1: 58.4000 (59.4065)  acc5: 80.0000 (82.0903)  time: 1.0683  data: 1.0492  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 1.7973 (1.9103)  acc1: 59.2000 (58.6927)  acc5: 81.6000 (81.6390)  time: 0.7439  data: 0.7244  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 1.9504 (1.9216)  acc1: 56.8000 (58.3680)  acc5: 79.2000 (81.1840)  time: 0.6244  data: 0.6043  max mem: 2905
Test: Total time: 0:00:49 (0.9824 s / it)
* Acc@1 59.092 Acc@5 81.518 loss 1.887
Accuracy of the model on the 50000 test images: 59.1%
Max accuracy: 59.63%
Epoch: [235]  [  0/625]  eta: 3:42:40  lr: 0.000510  min_lr: 0.000510  loss: 2.7276 (2.7276)  class_acc: 0.5664 (0.5664)  weight_decay: 0.0500 (0.0500)  time: 21.3775  data: 21.2518  max mem: 2905
Epoch: [235]  [200/625]  eta: 0:14:05  lr: 0.000505  min_lr: 0.000505  loss: 2.7797 (2.7879)  class_acc: 0.5742 (0.5741)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6321 (3.5089)  time: 1.8146  data: 1.4181  max mem: 2905
Epoch: [235]  [400/625]  eta: 0:07:19  lr: 0.000500  min_lr: 0.000500  loss: 2.8162 (2.7937)  class_acc: 0.5586 (0.5730)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.1491 (3.4641)  time: 1.9852  data: 1.6353  max mem: 2905
Epoch: [235]  [600/625]  eta: 0:00:48  lr: 0.000495  min_lr: 0.000495  loss: 2.8075 (2.7920)  class_acc: 0.5742 (0.5742)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.8011 (3.3297)  time: 1.9821  data: 1.1157  max mem: 2905
Epoch: [235]  [624/625]  eta: 0:00:01  lr: 0.000495  min_lr: 0.000495  loss: 2.8226 (2.7932)  class_acc: 0.5664 (0.5739)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5230 (3.3320)  time: 0.4779  data: 0.1769  max mem: 2905
Epoch: [235] Total time: 0:20:00 (1.9213 s / it)
Averaged stats: lr: 0.000495  min_lr: 0.000495  loss: 2.8226 (2.7872)  class_acc: 0.5664 (0.5753)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5230 (3.3320)
Test:  [ 0/50]  eta: 0:10:39  loss: 1.8610 (1.8610)  acc1: 63.2000 (63.2000)  acc5: 83.2000 (83.2000)  time: 12.7932  data: 12.7570  max mem: 2905
Test:  [10/50]  eta: 0:01:24  loss: 1.8463 (1.9009)  acc1: 56.8000 (58.0364)  acc5: 80.8000 (81.6727)  time: 2.1231  data: 2.1025  max mem: 2905
Test:  [20/50]  eta: 0:00:50  loss: 1.9796 (2.0586)  acc1: 54.4000 (55.1238)  acc5: 80.0000 (79.3143)  time: 1.1169  data: 1.0968  max mem: 2905
Test:  [30/50]  eta: 0:00:28  loss: 2.1496 (2.0463)  acc1: 55.2000 (55.5355)  acc5: 77.6000 (79.0710)  time: 1.0608  data: 1.0413  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 2.0426 (2.0689)  acc1: 56.0000 (55.3171)  acc5: 77.6000 (78.4390)  time: 0.6191  data: 0.5990  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 2.0900 (2.0753)  acc1: 54.4000 (55.1840)  acc5: 76.0000 (78.1760)  time: 0.6094  data: 0.5881  max mem: 2905
Test: Total time: 0:00:48 (0.9618 s / it)
* Acc@1 55.816 Acc@5 78.690 loss 2.041
Accuracy of the model on the 50000 test images: 55.8%
Max accuracy: 59.63%
Epoch: [236]  [  0/625]  eta: 3:41:12  lr: 0.000495  min_lr: 0.000495  loss: 2.6949 (2.6949)  class_acc: 0.5820 (0.5820)  weight_decay: 0.0500 (0.0500)  time: 21.2365  data: 16.8783  max mem: 2905
Epoch: [236]  [200/625]  eta: 0:13:55  lr: 0.000490  min_lr: 0.000490  loss: 2.7732 (2.7767)  class_acc: 0.5664 (0.5787)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.0168 (3.3288)  time: 1.9275  data: 0.9510  max mem: 2905
Epoch: [236]  [400/625]  eta: 0:07:15  lr: 0.000485  min_lr: 0.000485  loss: 2.7939 (2.7811)  class_acc: 0.5820 (0.5781)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4715 (3.2278)  time: 1.9014  data: 0.0238  max mem: 2905
Epoch: [236]  [600/625]  eta: 0:00:48  lr: 0.000481  min_lr: 0.000481  loss: 2.8003 (2.7817)  class_acc: 0.5703 (0.5774)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.9899 (3.2114)  time: 2.0740  data: 0.0666  max mem: 2905
Epoch: [236]  [624/625]  eta: 0:00:01  lr: 0.000480  min_lr: 0.000480  loss: 2.8196 (2.7836)  class_acc: 0.5742 (0.5770)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.2263 (3.2220)  time: 0.8521  data: 0.0015  max mem: 2905
Epoch: [236] Total time: 0:19:52 (1.9083 s / it)
Averaged stats: lr: 0.000480  min_lr: 0.000480  loss: 2.8196 (2.7873)  class_acc: 0.5742 (0.5756)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.2263 (3.2220)
Test:  [ 0/50]  eta: 0:10:38  loss: 1.7233 (1.7233)  acc1: 59.2000 (59.2000)  acc5: 84.8000 (84.8000)  time: 12.7606  data: 12.7294  max mem: 2905
Test:  [10/50]  eta: 0:01:24  loss: 1.6955 (1.6432)  acc1: 62.4000 (62.7636)  acc5: 81.6000 (83.7818)  time: 2.1201  data: 2.1006  max mem: 2905
Test:  [20/50]  eta: 0:00:50  loss: 1.8052 (1.8365)  acc1: 59.2000 (58.8571)  acc5: 81.6000 (81.7524)  time: 1.1204  data: 1.1022  max mem: 2905
Test:  [30/50]  eta: 0:00:29  loss: 1.9595 (1.8762)  acc1: 56.0000 (58.3484)  acc5: 81.6000 (81.1871)  time: 1.1416  data: 1.1236  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 1.9595 (1.9142)  acc1: 55.2000 (57.6000)  acc5: 80.8000 (80.5854)  time: 0.8884  data: 0.8701  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 1.9773 (1.9414)  acc1: 53.6000 (57.0240)  acc5: 78.4000 (80.3520)  time: 0.8351  data: 0.8158  max mem: 2905
Test: Total time: 0:00:53 (1.0796 s / it)
* Acc@1 58.618 Acc@5 81.202 loss 1.891
Accuracy of the model on the 50000 test images: 58.6%
Max accuracy: 59.63%
Epoch: [237]  [  0/625]  eta: 3:56:52  lr: 0.000480  min_lr: 0.000480  loss: 2.5955 (2.5955)  class_acc: 0.6055 (0.6055)  weight_decay: 0.0500 (0.0500)  time: 22.7404  data: 18.9484  max mem: 2905
Epoch: [237]  [200/625]  eta: 0:14:19  lr: 0.000475  min_lr: 0.000475  loss: 2.7488 (2.7758)  class_acc: 0.5742 (0.5772)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.0349 (3.4789)  time: 1.8228  data: 0.0007  max mem: 2905
Epoch: [237]  [400/625]  eta: 0:07:27  lr: 0.000471  min_lr: 0.000471  loss: 2.7498 (2.7823)  class_acc: 0.5820 (0.5761)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.0190 (3.5008)  time: 2.0332  data: 0.0007  max mem: 2905
Epoch: [237]  [600/625]  eta: 0:00:49  lr: 0.000466  min_lr: 0.000466  loss: 2.8107 (2.7830)  class_acc: 0.5742 (0.5766)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.9216 (3.4565)  time: 1.9514  data: 0.0007  max mem: 2905
Epoch: [237]  [624/625]  eta: 0:00:01  lr: 0.000466  min_lr: 0.000466  loss: 2.7838 (2.7839)  class_acc: 0.5703 (0.5763)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.5205 (3.4671)  time: 0.5957  data: 0.0020  max mem: 2905
Epoch: [237] Total time: 0:20:31 (1.9701 s / it)
Averaged stats: lr: 0.000466  min_lr: 0.000466  loss: 2.7838 (2.7830)  class_acc: 0.5703 (0.5761)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.5205 (3.4671)
Test:  [ 0/50]  eta: 0:10:47  loss: 1.9279 (1.9279)  acc1: 54.4000 (54.4000)  acc5: 84.0000 (84.0000)  time: 12.9461  data: 12.9142  max mem: 2905
Test:  [10/50]  eta: 0:01:26  loss: 1.7044 (1.6345)  acc1: 61.6000 (62.4000)  acc5: 84.8000 (85.0182)  time: 2.1507  data: 2.1309  max mem: 2905
Test:  [20/50]  eta: 0:00:50  loss: 1.7869 (1.8063)  acc1: 60.0000 (59.7714)  acc5: 82.4000 (83.2381)  time: 1.1073  data: 1.0890  max mem: 2905
Test:  [30/50]  eta: 0:00:29  loss: 1.9130 (1.8073)  acc1: 57.6000 (60.0774)  acc5: 82.4000 (82.9161)  time: 1.1184  data: 1.0984  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 1.7461 (1.8187)  acc1: 59.2000 (59.8829)  acc5: 83.2000 (82.6342)  time: 0.8932  data: 0.8735  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 1.7483 (1.8228)  acc1: 59.2000 (60.0160)  acc5: 81.6000 (82.3680)  time: 0.8634  data: 0.8441  max mem: 2905
Test: Total time: 0:00:53 (1.0758 s / it)
* Acc@1 61.060 Acc@5 82.848 loss 1.782
Accuracy of the model on the 50000 test images: 61.1%
Max accuracy: 61.06%
Epoch: [238]  [  0/625]  eta: 3:30:29  lr: 0.000466  min_lr: 0.000466  loss: 2.8227 (2.8227)  class_acc: 0.5859 (0.5859)  weight_decay: 0.0500 (0.0500)  time: 20.2075  data: 16.8635  max mem: 2905
Epoch: [238]  [200/625]  eta: 0:14:45  lr: 0.000461  min_lr: 0.000461  loss: 2.7656 (2.7678)  class_acc: 0.5742 (0.5817)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.9293 (3.2914)  time: 1.9695  data: 0.0738  max mem: 2905
Epoch: [238]  [400/625]  eta: 0:07:31  lr: 0.000456  min_lr: 0.000456  loss: 2.8005 (2.7783)  class_acc: 0.5625 (0.5786)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6832 (3.1466)  time: 1.9229  data: 0.0505  max mem: 2905
Epoch: [238]  [600/625]  eta: 0:00:50  lr: 0.000452  min_lr: 0.000452  loss: 2.7760 (2.7828)  class_acc: 0.5742 (0.5782)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.4062 (3.3167)  time: 1.9473  data: 0.0008  max mem: 2905
Epoch: [238]  [624/625]  eta: 0:00:01  lr: 0.000451  min_lr: 0.000451  loss: 2.7654 (2.7826)  class_acc: 0.5781 (0.5784)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.0902 (3.2842)  time: 0.8888  data: 0.0016  max mem: 2905
Epoch: [238] Total time: 0:20:29 (1.9672 s / it)
Averaged stats: lr: 0.000451  min_lr: 0.000451  loss: 2.7654 (2.7806)  class_acc: 0.5781 (0.5774)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.0902 (3.2842)
Test:  [ 0/50]  eta: 0:10:04  loss: 1.7888 (1.7888)  acc1: 58.4000 (58.4000)  acc5: 85.6000 (85.6000)  time: 12.0896  data: 12.0571  max mem: 2905
Test:  [10/50]  eta: 0:01:22  loss: 1.5926 (1.5937)  acc1: 65.6000 (64.1455)  acc5: 86.4000 (84.9455)  time: 2.0619  data: 2.0416  max mem: 2905
Test:  [20/50]  eta: 0:00:48  loss: 1.7535 (1.7433)  acc1: 60.8000 (61.3333)  acc5: 84.8000 (83.7714)  time: 1.0883  data: 1.0691  max mem: 2905
Test:  [30/50]  eta: 0:00:28  loss: 1.8361 (1.7455)  acc1: 59.2000 (61.5742)  acc5: 82.4000 (83.5097)  time: 1.0532  data: 1.0339  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 1.6893 (1.7610)  acc1: 61.6000 (61.5415)  acc5: 83.2000 (83.2390)  time: 0.7425  data: 0.7237  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 1.8629 (1.7830)  acc1: 57.6000 (60.7040)  acc5: 82.4000 (82.9440)  time: 0.6398  data: 0.6208  max mem: 2905
Test: Total time: 0:00:49 (0.9951 s / it)
* Acc@1 61.324 Acc@5 83.114 loss 1.754
Accuracy of the model on the 50000 test images: 61.3%
Max accuracy: 61.32%
Epoch: [239]  [  0/625]  eta: 3:38:04  lr: 0.000451  min_lr: 0.000451  loss: 2.7719 (2.7719)  class_acc: 0.6055 (0.6055)  weight_decay: 0.0500 (0.0500)  time: 20.9347  data: 20.6136  max mem: 2905
Epoch: [239]  [200/625]  eta: 0:14:35  lr: 0.000447  min_lr: 0.000447  loss: 2.7616 (2.7669)  class_acc: 0.5859 (0.5805)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.1448 (3.2851)  time: 1.8818  data: 0.0009  max mem: 2905
Epoch: [239]  [400/625]  eta: 0:07:30  lr: 0.000442  min_lr: 0.000442  loss: 2.7580 (2.7741)  class_acc: 0.5703 (0.5780)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.5243 (inf)  time: 1.9657  data: 0.0012  max mem: 2905
Epoch: [239]  [600/625]  eta: 0:00:50  lr: 0.000438  min_lr: 0.000438  loss: 2.7749 (2.7753)  class_acc: 0.5703 (0.5788)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6752 (inf)  time: 2.1050  data: 0.0008  max mem: 2905
Epoch: [239]  [624/625]  eta: 0:00:01  lr: 0.000437  min_lr: 0.000437  loss: 2.7535 (2.7757)  class_acc: 0.5781 (0.5788)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3334 (inf)  time: 1.0655  data: 0.0013  max mem: 2905
Epoch: [239] Total time: 0:20:30 (1.9687 s / it)
Averaged stats: lr: 0.000437  min_lr: 0.000437  loss: 2.7535 (2.7768)  class_acc: 0.5781 (0.5779)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3334 (inf)
Test:  [ 0/50]  eta: 0:10:40  loss: 1.5106 (1.5106)  acc1: 68.8000 (68.8000)  acc5: 88.0000 (88.0000)  time: 12.8081  data: 12.7803  max mem: 2905
Test:  [10/50]  eta: 0:01:20  loss: 1.5905 (1.5839)  acc1: 66.4000 (66.4727)  acc5: 87.2000 (86.1091)  time: 2.0068  data: 1.9859  max mem: 2905
Test:  [20/50]  eta: 0:00:46  loss: 1.7188 (1.7741)  acc1: 61.6000 (61.7524)  acc5: 84.0000 (83.5810)  time: 0.9711  data: 0.9520  max mem: 2905
Test:  [30/50]  eta: 0:00:25  loss: 1.8920 (1.7872)  acc1: 60.0000 (61.6000)  acc5: 82.4000 (83.1226)  time: 0.8914  data: 0.8730  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 1.7891 (1.8236)  acc1: 60.8000 (61.0146)  acc5: 82.4000 (82.2439)  time: 0.7275  data: 0.7084  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 1.7935 (1.8260)  acc1: 60.8000 (60.9280)  acc5: 84.0000 (82.3360)  time: 0.6560  data: 0.6364  max mem: 2905
Test: Total time: 0:00:51 (1.0203 s / it)
* Acc@1 60.740 Acc@5 82.674 loss 1.801
Accuracy of the model on the 50000 test images: 60.7%
Max accuracy: 61.32%
Epoch: [240]  [  0/625]  eta: 3:43:30  lr: 0.000437  min_lr: 0.000437  loss: 2.7879 (2.7879)  class_acc: 0.5938 (0.5938)  weight_decay: 0.0500 (0.0500)  time: 21.4568  data: 20.2205  max mem: 2905
Epoch: [240]  [200/625]  eta: 0:14:16  lr: 0.000433  min_lr: 0.000433  loss: 2.7671 (2.7668)  class_acc: 0.5742 (0.5787)  weight_decay: 0.0500 (0.0500)  grad_norm: 1.9119 (3.5169)  time: 1.8851  data: 0.2731  max mem: 2905
Epoch: [240]  [400/625]  eta: 0:07:26  lr: 0.000428  min_lr: 0.000428  loss: 2.7443 (2.7701)  class_acc: 0.5703 (0.5779)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.4075 (3.4219)  time: 2.1454  data: 0.8892  max mem: 2905
Epoch: [240]  [600/625]  eta: 0:00:49  lr: 0.000424  min_lr: 0.000424  loss: 2.7571 (2.7697)  class_acc: 0.5781 (0.5780)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.3776 (3.3400)  time: 2.0084  data: 1.1536  max mem: 2905
Epoch: [240]  [624/625]  eta: 0:00:01  lr: 0.000423  min_lr: 0.000423  loss: 2.7996 (2.7713)  class_acc: 0.5742 (0.5778)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.9011 (3.3337)  time: 0.9011  data: 0.2608  max mem: 2905
Epoch: [240] Total time: 0:20:10 (1.9369 s / it)
Averaged stats: lr: 0.000423  min_lr: 0.000423  loss: 2.7996 (2.7737)  class_acc: 0.5742 (0.5782)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.9011 (3.3337)
Test:  [ 0/50]  eta: 0:10:26  loss: 1.6511 (1.6511)  acc1: 61.6000 (61.6000)  acc5: 87.2000 (87.2000)  time: 12.5385  data: 12.4803  max mem: 2905
Test:  [10/50]  eta: 0:01:21  loss: 1.5726 (1.5675)  acc1: 64.0000 (65.1636)  acc5: 86.4000 (85.8909)  time: 2.0290  data: 2.0062  max mem: 2905
Test:  [20/50]  eta: 0:00:46  loss: 1.7784 (1.7399)  acc1: 62.4000 (61.5238)  acc5: 84.8000 (84.0000)  time: 0.9967  data: 0.9775  max mem: 2905
Test:  [30/50]  eta: 0:00:26  loss: 1.8981 (1.7510)  acc1: 57.6000 (61.3419)  acc5: 80.8000 (83.3806)  time: 0.9635  data: 0.9442  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 1.8026 (1.7835)  acc1: 57.6000 (60.9756)  acc5: 80.0000 (83.0829)  time: 0.6764  data: 0.6574  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 1.7653 (1.7831)  acc1: 59.2000 (60.8000)  acc5: 81.6000 (82.9440)  time: 0.6009  data: 0.5824  max mem: 2905
Test: Total time: 0:00:46 (0.9382 s / it)
* Acc@1 61.128 Acc@5 83.264 loss 1.761
Accuracy of the model on the 50000 test images: 61.1%
Max accuracy: 61.32%
Epoch: [241]  [  0/625]  eta: 3:22:33  lr: 0.000423  min_lr: 0.000423  loss: 2.5817 (2.5817)  class_acc: 0.6289 (0.6289)  weight_decay: 0.0500 (0.0500)  time: 19.4459  data: 19.2134  max mem: 2905
Epoch: [241]  [200/625]  eta: 0:13:46  lr: 0.000419  min_lr: 0.000419  loss: 2.7714 (2.7652)  class_acc: 0.5625 (0.5800)  weight_decay: 0.0500 (0.0500)  grad_norm: 4.1017 (3.8157)  time: 1.6421  data: 0.2635  max mem: 2905
Epoch: [241]  [400/625]  eta: 0:07:10  lr: 0.000415  min_lr: 0.000415  loss: 2.7636 (2.7750)  class_acc: 0.5781 (0.5781)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5308 (3.4547)  time: 1.8221  data: 0.0007  max mem: 2905
Epoch: [241]  [600/625]  eta: 0:00:48  lr: 0.000410  min_lr: 0.000410  loss: 2.7548 (2.7759)  class_acc: 0.5820 (0.5779)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6641 (3.5108)  time: 1.9151  data: 0.0011  max mem: 2905
Epoch: [241]  [624/625]  eta: 0:00:01  lr: 0.000410  min_lr: 0.000410  loss: 2.7937 (2.7759)  class_acc: 0.5742 (0.5780)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5352 (3.4761)  time: 0.9106  data: 0.0184  max mem: 2905
Epoch: [241] Total time: 0:19:41 (1.8903 s / it)
Averaged stats: lr: 0.000410  min_lr: 0.000410  loss: 2.7937 (2.7716)  class_acc: 0.5742 (0.5789)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5352 (3.4761)
Test:  [ 0/50]  eta: 0:10:59  loss: 1.7449 (1.7449)  acc1: 58.4000 (58.4000)  acc5: 84.0000 (84.0000)  time: 13.1914  data: 13.1631  max mem: 2905
Test:  [10/50]  eta: 0:01:34  loss: 1.6031 (1.6393)  acc1: 62.4000 (62.6909)  acc5: 85.6000 (84.5818)  time: 2.3723  data: 2.3519  max mem: 2905
Test:  [20/50]  eta: 0:00:53  loss: 1.7851 (1.7686)  acc1: 60.0000 (60.0000)  acc5: 83.2000 (82.9714)  time: 1.2280  data: 1.2090  max mem: 2905
Test:  [30/50]  eta: 0:00:30  loss: 1.8498 (1.7780)  acc1: 58.4000 (60.2581)  acc5: 81.6000 (82.6323)  time: 1.0776  data: 1.0579  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 1.7058 (1.7914)  acc1: 60.8000 (59.8439)  acc5: 81.6000 (82.2829)  time: 0.7644  data: 0.7430  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 1.7043 (1.7920)  acc1: 57.6000 (59.8080)  acc5: 82.4000 (82.2240)  time: 0.6467  data: 0.6240  max mem: 2905
Test: Total time: 0:00:53 (1.0745 s / it)
* Acc@1 60.942 Acc@5 82.952 loss 1.751
Accuracy of the model on the 50000 test images: 60.9%
Max accuracy: 61.32%
Epoch: [242]  [  0/625]  eta: 3:57:18  lr: 0.000410  min_lr: 0.000410  loss: 2.7184 (2.7184)  class_acc: 0.5742 (0.5742)  weight_decay: 0.0500 (0.0500)  time: 22.7819  data: 15.3494  max mem: 2905
Epoch: [242]  [200/625]  eta: 0:14:22  lr: 0.000405  min_lr: 0.000405  loss: 2.7780 (2.7539)  class_acc: 0.5859 (0.5825)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.8875 (3.0289)  time: 1.9559  data: 0.0013  max mem: 2905
Epoch: [242]  [400/625]  eta: 0:07:23  lr: 0.000401  min_lr: 0.000401  loss: 2.7109 (2.7539)  class_acc: 0.5898 (0.5812)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.2530 (3.1185)  time: 1.9582  data: 0.0008  max mem: 2905
Epoch: [242]  [600/625]  eta: 0:00:49  lr: 0.000397  min_lr: 0.000397  loss: 2.7970 (2.7612)  class_acc: 0.5703 (0.5806)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.0874 (3.2829)  time: 2.1660  data: 0.0134  max mem: 2905
Epoch: [242]  [624/625]  eta: 0:00:01  lr: 0.000396  min_lr: 0.000396  loss: 2.8043 (2.7616)  class_acc: 0.5703 (0.5803)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5360 (3.2674)  time: 0.7304  data: 0.0017  max mem: 2905
Epoch: [242] Total time: 0:20:06 (1.9296 s / it)
Averaged stats: lr: 0.000396  min_lr: 0.000396  loss: 2.8043 (2.7708)  class_acc: 0.5703 (0.5793)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5360 (3.2674)
Test:  [ 0/50]  eta: 0:10:53  loss: 1.7314 (1.7314)  acc1: 58.4000 (58.4000)  acc5: 85.6000 (85.6000)  time: 13.0795  data: 13.0457  max mem: 2905
Test:  [10/50]  eta: 0:01:21  loss: 1.6015 (1.5680)  acc1: 64.8000 (63.9273)  acc5: 86.4000 (85.6727)  time: 2.0291  data: 2.0077  max mem: 2905
Test:  [20/50]  eta: 0:00:45  loss: 1.6664 (1.6840)  acc1: 62.4000 (62.7429)  acc5: 84.8000 (84.4952)  time: 0.9548  data: 0.9347  max mem: 2905
Test:  [30/50]  eta: 0:00:26  loss: 1.7986 (1.7048)  acc1: 60.8000 (62.2452)  acc5: 83.2000 (84.2065)  time: 0.9521  data: 0.9315  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 1.7986 (1.7233)  acc1: 58.4000 (61.5610)  acc5: 82.4000 (83.6488)  time: 0.8802  data: 0.8601  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 1.7420 (1.7317)  acc1: 58.4000 (61.5360)  acc5: 82.4000 (83.4400)  time: 0.6336  data: 0.6143  max mem: 2905
Test: Total time: 0:00:53 (1.0720 s / it)
* Acc@1 62.014 Acc@5 83.670 loss 1.703
Accuracy of the model on the 50000 test images: 62.0%
Max accuracy: 62.01%
Epoch: [243]  [  0/625]  eta: 3:40:31  lr: 0.000396  min_lr: 0.000396  loss: 2.7409 (2.7409)  class_acc: 0.5625 (0.5625)  weight_decay: 0.0500 (0.0500)  time: 21.1696  data: 17.5611  max mem: 2905
Epoch: [243]  [200/625]  eta: 0:14:28  lr: 0.000392  min_lr: 0.000392  loss: 2.7529 (2.7562)  class_acc: 0.5859 (0.5865)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3311 (3.4300)  time: 1.7437  data: 0.0007  max mem: 2905
Epoch: [243]  [400/625]  eta: 0:07:31  lr: 0.000388  min_lr: 0.000388  loss: 2.7159 (2.7591)  class_acc: 0.5898 (0.5845)  weight_decay: 0.0500 (0.0500)  grad_norm: 4.4367 (3.3550)  time: 1.9394  data: 0.0007  max mem: 2905
Epoch: [243]  [600/625]  eta: 0:00:50  lr: 0.000383  min_lr: 0.000383  loss: 2.7774 (2.7662)  class_acc: 0.5742 (0.5817)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.1733 (3.3561)  time: 2.1701  data: 0.0007  max mem: 2905
Epoch: [243]  [624/625]  eta: 0:00:01  lr: 0.000383  min_lr: 0.000383  loss: 2.7374 (2.7669)  class_acc: 0.5664 (0.5813)  weight_decay: 0.0500 (0.0500)  grad_norm: 4.5450 (3.4273)  time: 0.8327  data: 0.0014  max mem: 2905
Epoch: [243] Total time: 0:20:22 (1.9557 s / it)
Averaged stats: lr: 0.000383  min_lr: 0.000383  loss: 2.7374 (2.7680)  class_acc: 0.5664 (0.5799)  weight_decay: 0.0500 (0.0500)  grad_norm: 4.5450 (3.4273)
Test:  [ 0/50]  eta: 0:10:31  loss: 1.7137 (1.7137)  acc1: 60.8000 (60.8000)  acc5: 84.0000 (84.0000)  time: 12.6229  data: 12.5925  max mem: 2905
Test:  [10/50]  eta: 0:01:15  loss: 1.5407 (1.6163)  acc1: 63.2000 (62.8364)  acc5: 84.0000 (84.7273)  time: 1.8877  data: 1.8658  max mem: 2905
Test:  [20/50]  eta: 0:00:39  loss: 1.7987 (1.7814)  acc1: 58.4000 (59.7714)  acc5: 83.2000 (82.8191)  time: 0.7564  data: 0.7356  max mem: 2905
Test:  [30/50]  eta: 0:00:25  loss: 1.8556 (1.7868)  acc1: 57.6000 (59.9484)  acc5: 81.6000 (82.9419)  time: 0.9324  data: 0.9128  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 1.8156 (1.8131)  acc1: 58.4000 (59.7463)  acc5: 82.4000 (82.1463)  time: 0.8742  data: 0.8544  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 1.7923 (1.8212)  acc1: 58.4000 (59.6800)  acc5: 82.4000 (82.2400)  time: 0.4421  data: 0.4222  max mem: 2905
Test: Total time: 0:00:48 (0.9680 s / it)
* Acc@1 60.508 Acc@5 82.950 loss 1.777
Accuracy of the model on the 50000 test images: 60.5%
Max accuracy: 62.01%
Epoch: [244]  [  0/625]  eta: 3:50:23  lr: 0.000383  min_lr: 0.000383  loss: 2.6059 (2.6059)  class_acc: 0.6172 (0.6172)  weight_decay: 0.0500 (0.0500)  time: 22.1176  data: 21.7752  max mem: 2905
Epoch: [244]  [200/625]  eta: 0:14:16  lr: 0.000379  min_lr: 0.000379  loss: 2.7288 (2.7459)  class_acc: 0.5938 (0.5829)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.0239 (3.3966)  time: 1.8827  data: 1.2112  max mem: 2905
Epoch: [244]  [400/625]  eta: 0:07:21  lr: 0.000374  min_lr: 0.000374  loss: 2.7790 (2.7590)  class_acc: 0.5898 (0.5818)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.4928 (3.4304)  time: 1.8949  data: 1.5867  max mem: 2905
Epoch: [244]  [600/625]  eta: 0:00:48  lr: 0.000370  min_lr: 0.000370  loss: 2.7377 (2.7632)  class_acc: 0.5898 (0.5819)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4940 (3.5859)  time: 1.6860  data: 0.2579  max mem: 2905
Epoch: [244]  [624/625]  eta: 0:00:01  lr: 0.000370  min_lr: 0.000370  loss: 2.7756 (2.7630)  class_acc: 0.5859 (0.5818)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.9518 (3.5723)  time: 0.6657  data: 0.0643  max mem: 2905
Epoch: [244] Total time: 0:19:54 (1.9111 s / it)
Averaged stats: lr: 0.000370  min_lr: 0.000370  loss: 2.7756 (2.7649)  class_acc: 0.5859 (0.5808)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.9518 (3.5723)
Test:  [ 0/50]  eta: 0:10:16  loss: 1.7577 (1.7577)  acc1: 63.2000 (63.2000)  acc5: 84.8000 (84.8000)  time: 12.3346  data: 12.2949  max mem: 2905
Test:  [10/50]  eta: 0:01:27  loss: 1.5378 (1.5262)  acc1: 67.2000 (66.6909)  acc5: 84.8000 (86.2545)  time: 2.1813  data: 2.1592  max mem: 2905
Test:  [20/50]  eta: 0:00:47  loss: 1.6097 (1.7092)  acc1: 61.6000 (61.8286)  acc5: 84.8000 (84.0381)  time: 1.0442  data: 1.0246  max mem: 2905
Test:  [30/50]  eta: 0:00:25  loss: 1.8465 (1.7113)  acc1: 58.4000 (61.8065)  acc5: 84.8000 (84.2065)  time: 0.7915  data: 0.7728  max mem: 2905
Test:  [40/50]  eta: 0:00:10  loss: 1.7804 (1.7387)  acc1: 58.4000 (61.3854)  acc5: 84.8000 (83.7073)  time: 0.4864  data: 0.4680  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 1.7521 (1.7475)  acc1: 58.4000 (61.2640)  acc5: 83.2000 (83.5040)  time: 0.3836  data: 0.3635  max mem: 2905
Test: Total time: 0:00:45 (0.9047 s / it)
* Acc@1 61.934 Acc@5 83.588 loss 1.710
Accuracy of the model on the 50000 test images: 61.9%
Max accuracy: 62.01%
Epoch: [245]  [  0/625]  eta: 3:25:52  lr: 0.000370  min_lr: 0.000370  loss: 2.7690 (2.7690)  class_acc: 0.5938 (0.5938)  weight_decay: 0.0500 (0.0500)  time: 19.7643  data: 19.2677  max mem: 2905
Epoch: [245]  [200/625]  eta: 0:14:06  lr: 0.000366  min_lr: 0.000366  loss: 2.7916 (2.7665)  class_acc: 0.5820 (0.5796)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.2291 (3.4862)  time: 2.0710  data: 0.0008  max mem: 2905
Epoch: [245]  [400/625]  eta: 0:07:19  lr: 0.000362  min_lr: 0.000362  loss: 2.7396 (2.7614)  class_acc: 0.5820 (0.5812)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.0495 (3.4060)  time: 2.0380  data: 0.0008  max mem: 2905
Epoch: [245]  [600/625]  eta: 0:00:48  lr: 0.000357  min_lr: 0.000357  loss: 2.7622 (2.7639)  class_acc: 0.5859 (0.5810)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7053 (3.4730)  time: 2.1348  data: 0.0333  max mem: 2905
Epoch: [245]  [624/625]  eta: 0:00:01  lr: 0.000357  min_lr: 0.000357  loss: 2.7437 (2.7637)  class_acc: 0.5781 (0.5809)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7557 (3.4413)  time: 0.7689  data: 0.0091  max mem: 2905
Epoch: [245] Total time: 0:19:55 (1.9122 s / it)
Averaged stats: lr: 0.000357  min_lr: 0.000357  loss: 2.7437 (2.7610)  class_acc: 0.5781 (0.5813)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7557 (3.4413)
Test:  [ 0/50]  eta: 0:10:08  loss: 1.7890 (1.7890)  acc1: 58.4000 (58.4000)  acc5: 83.2000 (83.2000)  time: 12.1649  data: 12.1250  max mem: 2905
Test:  [10/50]  eta: 0:01:21  loss: 1.5829 (1.5946)  acc1: 64.0000 (64.2909)  acc5: 87.2000 (85.4546)  time: 2.0292  data: 2.0075  max mem: 2905
Test:  [20/50]  eta: 0:00:48  loss: 1.7567 (1.7307)  acc1: 60.8000 (61.3714)  acc5: 84.0000 (83.7714)  time: 1.0730  data: 1.0526  max mem: 2905
Test:  [30/50]  eta: 0:00:28  loss: 1.7886 (1.7307)  acc1: 59.2000 (60.9290)  acc5: 82.4000 (83.4323)  time: 1.1025  data: 1.0820  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 1.7137 (1.7667)  acc1: 58.4000 (60.2927)  acc5: 82.4000 (82.7707)  time: 0.9474  data: 0.9271  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 1.8086 (1.7832)  acc1: 58.4000 (59.9520)  acc5: 82.4000 (82.6240)  time: 0.8457  data: 0.8251  max mem: 2905
Test: Total time: 0:00:56 (1.1220 s / it)
* Acc@1 60.762 Acc@5 82.960 loss 1.757
Accuracy of the model on the 50000 test images: 60.8%
Max accuracy: 62.01%
Epoch: [246]  [  0/625]  eta: 3:23:36  lr: 0.000357  min_lr: 0.000357  loss: 2.6648 (2.6648)  class_acc: 0.5625 (0.5625)  weight_decay: 0.0500 (0.0500)  time: 19.5465  data: 19.2392  max mem: 2905
Epoch: [246]  [200/625]  eta: 0:13:50  lr: 0.000353  min_lr: 0.000353  loss: 2.7346 (2.7542)  class_acc: 0.5781 (0.5838)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.4239 (3.1977)  time: 1.8317  data: 0.0006  max mem: 2905
Epoch: [246]  [400/625]  eta: 0:07:10  lr: 0.000349  min_lr: 0.000349  loss: 2.7236 (2.7589)  class_acc: 0.5859 (0.5829)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.0873 (3.3556)  time: 1.8041  data: 0.0007  max mem: 2905
Epoch: [246]  [600/625]  eta: 0:00:47  lr: 0.000345  min_lr: 0.000345  loss: 2.7883 (2.7603)  class_acc: 0.5742 (0.5825)  weight_decay: 0.0500 (0.0500)  grad_norm: 5.4175 (inf)  time: 1.8969  data: 0.8910  max mem: 2905
Epoch: [246]  [624/625]  eta: 0:00:01  lr: 0.000344  min_lr: 0.000344  loss: 2.7428 (2.7597)  class_acc: 0.5820 (0.5827)  weight_decay: 0.0500 (0.0500)  grad_norm: 5.7122 (inf)  time: 0.6722  data: 0.3728  max mem: 2905
Epoch: [246] Total time: 0:19:39 (1.8868 s / it)
Averaged stats: lr: 0.000344  min_lr: 0.000344  loss: 2.7428 (2.7590)  class_acc: 0.5820 (0.5821)  weight_decay: 0.0500 (0.0500)  grad_norm: 5.7122 (inf)
Test:  [ 0/50]  eta: 0:10:12  loss: 1.6667 (1.6667)  acc1: 62.4000 (62.4000)  acc5: 86.4000 (86.4000)  time: 12.2585  data: 12.2244  max mem: 2905
Test:  [10/50]  eta: 0:01:16  loss: 1.4992 (1.5292)  acc1: 65.6000 (66.1091)  acc5: 86.4000 (85.6727)  time: 1.9172  data: 1.8927  max mem: 2905
Test:  [20/50]  eta: 0:00:42  loss: 1.6921 (1.7046)  acc1: 64.0000 (62.4762)  acc5: 83.2000 (83.5429)  time: 0.8749  data: 0.8537  max mem: 2905
Test:  [30/50]  eta: 0:00:24  loss: 1.8057 (1.7293)  acc1: 57.6000 (61.5484)  acc5: 81.6000 (83.2258)  time: 0.8376  data: 0.8178  max mem: 2905
Test:  [40/50]  eta: 0:00:10  loss: 1.7670 (1.7485)  acc1: 57.6000 (61.1707)  acc5: 82.4000 (83.0634)  time: 0.6940  data: 0.6742  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 1.7713 (1.7519)  acc1: 58.4000 (60.9760)  acc5: 80.8000 (82.8960)  time: 0.5623  data: 0.5435  max mem: 2905
Test: Total time: 0:00:46 (0.9349 s / it)
* Acc@1 61.762 Acc@5 83.206 loss 1.718
Accuracy of the model on the 50000 test images: 61.8%
Max accuracy: 62.01%
Epoch: [247]  [  0/625]  eta: 3:26:35  lr: 0.000344  min_lr: 0.000344  loss: 2.8539 (2.8539)  class_acc: 0.5703 (0.5703)  weight_decay: 0.0500 (0.0500)  time: 19.8328  data: 17.3093  max mem: 2905
Epoch: [247]  [200/625]  eta: 0:13:45  lr: 0.000340  min_lr: 0.000340  loss: 2.7122 (2.7429)  class_acc: 0.5977 (0.5877)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1684 (2.9796)  time: 1.9026  data: 0.0481  max mem: 2905
Epoch: [247]  [400/625]  eta: 0:07:10  lr: 0.000336  min_lr: 0.000336  loss: 2.7419 (2.7444)  class_acc: 0.5781 (0.5859)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.9002 (3.2321)  time: 1.7504  data: 0.0677  max mem: 2905
Epoch: [247]  [600/625]  eta: 0:00:48  lr: 0.000332  min_lr: 0.000332  loss: 2.7652 (2.7500)  class_acc: 0.5703 (0.5845)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.4353 (3.2385)  time: 1.9152  data: 0.0176  max mem: 2905
Epoch: [247]  [624/625]  eta: 0:00:01  lr: 0.000332  min_lr: 0.000332  loss: 2.7836 (2.7508)  class_acc: 0.5938 (0.5845)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.7755 (3.2782)  time: 0.7040  data: 0.0314  max mem: 2905
Epoch: [247] Total time: 0:19:51 (1.9058 s / it)
Averaged stats: lr: 0.000332  min_lr: 0.000332  loss: 2.7836 (2.7552)  class_acc: 0.5938 (0.5830)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.7755 (3.2782)
Test:  [ 0/50]  eta: 0:10:14  loss: 1.7678 (1.7678)  acc1: 60.0000 (60.0000)  acc5: 86.4000 (86.4000)  time: 12.2941  data: 12.2687  max mem: 2905
Test:  [10/50]  eta: 0:01:25  loss: 1.6385 (1.6347)  acc1: 63.2000 (63.5636)  acc5: 85.6000 (84.7273)  time: 2.1481  data: 2.1265  max mem: 2905
Test:  [20/50]  eta: 0:00:50  loss: 1.7778 (1.8048)  acc1: 60.0000 (59.7333)  acc5: 83.2000 (83.0857)  time: 1.1696  data: 1.1496  max mem: 2905
Test:  [30/50]  eta: 0:00:29  loss: 1.8623 (1.7954)  acc1: 59.2000 (60.1806)  acc5: 83.2000 (82.9936)  time: 1.1259  data: 1.1068  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 1.7402 (1.8095)  acc1: 60.0000 (59.8049)  acc5: 81.6000 (82.4195)  time: 0.7296  data: 0.7103  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 1.7686 (1.8017)  acc1: 60.8000 (60.1120)  acc5: 81.6000 (82.2560)  time: 0.6405  data: 0.6197  max mem: 2905
Test: Total time: 0:00:51 (1.0276 s / it)
* Acc@1 60.984 Acc@5 82.372 loss 1.780
Accuracy of the model on the 50000 test images: 61.0%
Max accuracy: 62.01%
Epoch: [248]  [  0/625]  eta: 3:36:18  lr: 0.000332  min_lr: 0.000332  loss: 2.7000 (2.7000)  class_acc: 0.6133 (0.6133)  weight_decay: 0.0500 (0.0500)  time: 20.7662  data: 18.4724  max mem: 2905
Epoch: [248]  [200/625]  eta: 0:14:37  lr: 0.000328  min_lr: 0.000328  loss: 2.7880 (2.7437)  class_acc: 0.5820 (0.5872)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7727 (3.2098)  time: 1.7260  data: 0.0006  max mem: 2905
Epoch: [248]  [400/625]  eta: 0:07:41  lr: 0.000324  min_lr: 0.000324  loss: 2.7056 (2.7480)  class_acc: 0.5820 (0.5846)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.9931 (3.2920)  time: 2.1214  data: 0.0006  max mem: 2905
Epoch: [248]  [600/625]  eta: 0:00:51  lr: 0.000320  min_lr: 0.000320  loss: 2.7750 (2.7493)  class_acc: 0.5742 (0.5833)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.8891 (3.2853)  time: 2.0283  data: 0.0006  max mem: 2905
Epoch: [248]  [624/625]  eta: 0:00:01  lr: 0.000320  min_lr: 0.000320  loss: 2.7426 (2.7486)  class_acc: 0.5820 (0.5834)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.0953 (3.2853)  time: 0.8414  data: 0.0013  max mem: 2905
Epoch: [248] Total time: 0:20:48 (1.9982 s / it)
Averaged stats: lr: 0.000320  min_lr: 0.000320  loss: 2.7426 (2.7526)  class_acc: 0.5820 (0.5836)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.0953 (3.2853)
Test:  [ 0/50]  eta: 0:09:00  loss: 1.5439 (1.5439)  acc1: 60.8000 (60.8000)  acc5: 86.4000 (86.4000)  time: 10.8025  data: 10.7716  max mem: 2905
Test:  [10/50]  eta: 0:01:18  loss: 1.4529 (1.4601)  acc1: 66.4000 (67.8545)  acc5: 88.0000 (87.3455)  time: 1.9718  data: 1.9526  max mem: 2905
Test:  [20/50]  eta: 0:00:48  loss: 1.6206 (1.6176)  acc1: 63.2000 (63.8857)  acc5: 85.6000 (85.4857)  time: 1.1564  data: 1.1389  max mem: 2905
Test:  [30/50]  eta: 0:00:29  loss: 1.7389 (1.6359)  acc1: 59.2000 (63.1742)  acc5: 84.0000 (84.8774)  time: 1.1676  data: 1.1490  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 1.7198 (1.6502)  acc1: 62.4000 (63.0244)  acc5: 82.4000 (84.5463)  time: 0.8768  data: 0.8564  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 1.7108 (1.6633)  acc1: 61.6000 (62.5760)  acc5: 83.2000 (84.3840)  time: 0.7465  data: 0.7260  max mem: 2905
Test: Total time: 0:00:51 (1.0378 s / it)
* Acc@1 63.404 Acc@5 84.742 loss 1.627
Accuracy of the model on the 50000 test images: 63.4%
Max accuracy: 63.40%
Epoch: [249]  [  0/625]  eta: 3:27:44  lr: 0.000320  min_lr: 0.000320  loss: 2.8956 (2.8956)  class_acc: 0.5352 (0.5352)  weight_decay: 0.0500 (0.0500)  time: 19.9438  data: 19.8201  max mem: 2905
Epoch: [249]  [200/625]  eta: 0:14:48  lr: 0.000316  min_lr: 0.000316  loss: 2.6596 (2.7379)  class_acc: 0.5977 (0.5869)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.8069 (3.5533)  time: 2.1266  data: 0.0007  max mem: 2905
Epoch: [249]  [400/625]  eta: 0:07:38  lr: 0.000312  min_lr: 0.000312  loss: 2.7756 (2.7489)  class_acc: 0.5820 (0.5852)  weight_decay: 0.0500 (0.0500)  grad_norm: 4.6658 (3.7656)  time: 2.0893  data: 0.0006  max mem: 2905
Epoch: [249]  [600/625]  eta: 0:00:50  lr: 0.000308  min_lr: 0.000308  loss: 2.7614 (2.7504)  class_acc: 0.5820 (0.5845)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7433 (3.7822)  time: 1.9091  data: 0.0006  max mem: 2905
Epoch: [249]  [624/625]  eta: 0:00:01  lr: 0.000308  min_lr: 0.000308  loss: 2.7108 (2.7491)  class_acc: 0.5898 (0.5848)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.8314 (3.7907)  time: 0.8130  data: 0.0109  max mem: 2905
Epoch: [249] Total time: 0:20:43 (1.9895 s / it)
Averaged stats: lr: 0.000308  min_lr: 0.000308  loss: 2.7108 (2.7496)  class_acc: 0.5898 (0.5843)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.8314 (3.7907)
Test:  [ 0/50]  eta: 0:09:41  loss: 1.7079 (1.7079)  acc1: 61.6000 (61.6000)  acc5: 83.2000 (83.2000)  time: 11.6289  data: 11.5878  max mem: 2905
Test:  [10/50]  eta: 0:01:20  loss: 1.5484 (1.5617)  acc1: 64.0000 (65.4545)  acc5: 84.8000 (84.6546)  time: 2.0162  data: 1.9944  max mem: 2905
Test:  [20/50]  eta: 0:00:48  loss: 1.7695 (1.7363)  acc1: 60.8000 (61.2191)  acc5: 82.4000 (82.8952)  time: 1.1250  data: 1.1055  max mem: 2905
Test:  [30/50]  eta: 0:00:29  loss: 1.7879 (1.7272)  acc1: 59.2000 (61.3419)  acc5: 82.4000 (83.2258)  time: 1.1621  data: 1.1432  max mem: 2905
Test:  [40/50]  eta: 0:00:13  loss: 1.7082 (1.7476)  acc1: 60.8000 (61.1902)  acc5: 83.2000 (82.9073)  time: 0.9970  data: 0.9781  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 1.6476 (1.7471)  acc1: 59.2000 (61.1200)  acc5: 83.2000 (82.8960)  time: 0.8783  data: 0.8599  max mem: 2905
Test: Total time: 0:00:55 (1.1118 s / it)
* Acc@1 61.994 Acc@5 83.418 loss 1.715
Accuracy of the model on the 50000 test images: 62.0%
Max accuracy: 63.40%
Epoch: [250]  [  0/625]  eta: 3:50:13  lr: 0.000307  min_lr: 0.000307  loss: 2.5409 (2.5409)  class_acc: 0.6289 (0.6289)  weight_decay: 0.0500 (0.0500)  time: 22.1023  data: 18.4403  max mem: 2905
Epoch: [250]  [200/625]  eta: 0:14:38  lr: 0.000304  min_lr: 0.000304  loss: 2.7303 (2.7308)  class_acc: 0.5742 (0.5875)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6854 (3.1862)  time: 1.9584  data: 0.0009  max mem: 2905
Epoch: [250]  [400/625]  eta: 0:07:35  lr: 0.000300  min_lr: 0.000300  loss: 2.7362 (2.7350)  class_acc: 0.5898 (0.5861)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6315 (3.2845)  time: 2.0521  data: 0.0197  max mem: 2905
Epoch: [250]  [600/625]  eta: 0:00:50  lr: 0.000296  min_lr: 0.000296  loss: 2.7671 (2.7396)  class_acc: 0.5664 (0.5857)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.9016 (3.3349)  time: 2.0908  data: 0.0007  max mem: 2905
Epoch: [250]  [624/625]  eta: 0:00:01  lr: 0.000296  min_lr: 0.000296  loss: 2.7524 (2.7401)  class_acc: 0.5859 (0.5855)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.9613 (3.3660)  time: 0.8337  data: 0.0016  max mem: 2905
Epoch: [250] Total time: 0:20:34 (1.9759 s / it)
Averaged stats: lr: 0.000296  min_lr: 0.000296  loss: 2.7524 (2.7462)  class_acc: 0.5859 (0.5850)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.9613 (3.3660)
Test:  [ 0/50]  eta: 0:08:56  loss: 1.7887 (1.7887)  acc1: 58.4000 (58.4000)  acc5: 85.6000 (85.6000)  time: 10.7376  data: 10.7108  max mem: 2905
Test:  [10/50]  eta: 0:01:00  loss: 1.5495 (1.5564)  acc1: 66.4000 (66.1091)  acc5: 85.6000 (85.8909)  time: 1.5227  data: 1.5022  max mem: 2905
Test:  [20/50]  eta: 0:00:39  loss: 1.6848 (1.7118)  acc1: 62.4000 (62.7048)  acc5: 83.2000 (84.0381)  time: 0.8385  data: 0.8184  max mem: 2905
Test:  [30/50]  eta: 0:00:23  loss: 1.8307 (1.7295)  acc1: 59.2000 (62.1936)  acc5: 81.6000 (83.4323)  time: 1.0025  data: 0.9827  max mem: 2905
Test:  [40/50]  eta: 0:00:10  loss: 1.7892 (1.7399)  acc1: 59.2000 (62.0683)  acc5: 82.4000 (83.0829)  time: 0.7374  data: 0.7183  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 1.7649 (1.7478)  acc1: 59.2000 (61.7120)  acc5: 81.6000 (82.7360)  time: 0.4111  data: 0.3928  max mem: 2905
Test: Total time: 0:00:45 (0.9034 s / it)
* Acc@1 62.198 Acc@5 83.376 loss 1.706
Accuracy of the model on the 50000 test images: 62.2%
Max accuracy: 63.40%
Epoch: [251]  [  0/625]  eta: 3:37:07  lr: 0.000296  min_lr: 0.000296  loss: 2.8539 (2.8539)  class_acc: 0.5430 (0.5430)  weight_decay: 0.0500 (0.0500)  time: 20.8443  data: 17.5066  max mem: 2905
Epoch: [251]  [200/625]  eta: 0:14:30  lr: 0.000292  min_lr: 0.000292  loss: 2.7129 (2.7388)  class_acc: 0.5938 (0.5871)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.8670 (3.3793)  time: 2.0413  data: 0.0007  max mem: 2905
Epoch: [251]  [400/625]  eta: 0:07:29  lr: 0.000288  min_lr: 0.000288  loss: 2.7485 (2.7504)  class_acc: 0.5938 (0.5842)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4842 (3.3045)  time: 1.9685  data: 0.0007  max mem: 2905
Epoch: [251]  [600/625]  eta: 0:00:50  lr: 0.000284  min_lr: 0.000284  loss: 2.7451 (2.7488)  class_acc: 0.5625 (0.5844)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5057 (3.2126)  time: 2.0468  data: 0.0742  max mem: 2905
Epoch: [251]  [624/625]  eta: 0:00:01  lr: 0.000284  min_lr: 0.000284  loss: 2.7682 (2.7501)  class_acc: 0.5703 (0.5840)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6611 (3.2146)  time: 1.0430  data: 0.0038  max mem: 2905
Epoch: [251] Total time: 0:20:44 (1.9906 s / it)
Averaged stats: lr: 0.000284  min_lr: 0.000284  loss: 2.7682 (2.7443)  class_acc: 0.5703 (0.5854)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6611 (3.2146)
Test:  [ 0/50]  eta: 0:10:06  loss: 1.6305 (1.6305)  acc1: 62.4000 (62.4000)  acc5: 84.8000 (84.8000)  time: 12.1314  data: 12.0983  max mem: 2905
Test:  [10/50]  eta: 0:01:19  loss: 1.5044 (1.5755)  acc1: 66.4000 (66.5455)  acc5: 86.4000 (85.6727)  time: 1.9828  data: 1.9625  max mem: 2905
Test:  [20/50]  eta: 0:00:50  loss: 1.7115 (1.7187)  acc1: 62.4000 (63.0095)  acc5: 84.0000 (84.2667)  time: 1.1436  data: 1.1240  max mem: 2905
Test:  [30/50]  eta: 0:00:31  loss: 1.7853 (1.7146)  acc1: 60.8000 (63.3806)  acc5: 82.4000 (83.9484)  time: 1.3518  data: 1.3296  max mem: 2905
Test:  [40/50]  eta: 0:00:13  loss: 1.6920 (1.7279)  acc1: 63.2000 (62.7317)  acc5: 83.2000 (83.7659)  time: 1.0631  data: 1.0414  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 1.6563 (1.7186)  acc1: 63.2000 (62.9760)  acc5: 84.8000 (84.0000)  time: 0.7976  data: 0.7785  max mem: 2905
Test: Total time: 0:00:58 (1.1614 s / it)
* Acc@1 63.504 Acc@5 84.526 loss 1.685
Accuracy of the model on the 50000 test images: 63.5%
Max accuracy: 63.50%
Epoch: [252]  [  0/625]  eta: 3:54:36  lr: 0.000284  min_lr: 0.000284  loss: 3.0420 (3.0420)  class_acc: 0.5469 (0.5469)  weight_decay: 0.0500 (0.0500)  time: 22.5231  data: 18.5886  max mem: 2905
Epoch: [252]  [200/625]  eta: 0:15:02  lr: 0.000280  min_lr: 0.000280  loss: 2.7254 (2.7393)  class_acc: 0.5742 (0.5848)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.9925 (3.4662)  time: 2.2164  data: 0.8763  max mem: 2905
Epoch: [252]  [400/625]  eta: 0:07:33  lr: 0.000277  min_lr: 0.000277  loss: 2.7541 (2.7445)  class_acc: 0.5742 (0.5846)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.7114 (3.6017)  time: 2.0685  data: 0.0007  max mem: 2905
Epoch: [252]  [600/625]  eta: 0:00:49  lr: 0.000273  min_lr: 0.000273  loss: 2.6950 (2.7407)  class_acc: 0.5938 (0.5863)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.8804 (3.4904)  time: 1.8444  data: 0.0006  max mem: 2905
Epoch: [252]  [624/625]  eta: 0:00:01  lr: 0.000273  min_lr: 0.000273  loss: 2.7226 (2.7408)  class_acc: 0.5781 (0.5862)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.1792 (3.5127)  time: 0.7131  data: 0.0014  max mem: 2905
Epoch: [252] Total time: 0:20:29 (1.9672 s / it)
Averaged stats: lr: 0.000273  min_lr: 0.000273  loss: 2.7226 (2.7406)  class_acc: 0.5781 (0.5862)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.1792 (3.5127)
Test:  [ 0/50]  eta: 0:10:37  loss: 1.6181 (1.6181)  acc1: 57.6000 (57.6000)  acc5: 88.8000 (88.8000)  time: 12.7449  data: 12.7117  max mem: 2905
Test:  [10/50]  eta: 0:01:25  loss: 1.3761 (1.3977)  acc1: 68.8000 (68.2182)  acc5: 88.0000 (88.3636)  time: 2.1367  data: 2.1169  max mem: 2905
Test:  [20/50]  eta: 0:00:50  loss: 1.5155 (1.5729)  acc1: 63.2000 (64.4191)  acc5: 85.6000 (86.2095)  time: 1.1292  data: 1.1090  max mem: 2905
Test:  [30/50]  eta: 0:00:29  loss: 1.6948 (1.5814)  acc1: 62.4000 (64.5936)  acc5: 84.8000 (85.7290)  time: 1.1339  data: 1.1134  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 1.6002 (1.6103)  acc1: 63.2000 (64.0390)  acc5: 84.8000 (85.1707)  time: 0.8725  data: 0.8520  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 1.5257 (1.6124)  acc1: 64.0000 (64.0480)  acc5: 84.8000 (85.1360)  time: 0.8463  data: 0.8264  max mem: 2905
Test: Total time: 0:00:53 (1.0719 s / it)
* Acc@1 64.746 Acc@5 85.416 loss 1.572
Accuracy of the model on the 50000 test images: 64.7%
Max accuracy: 64.75%
Epoch: [253]  [  0/625]  eta: 3:44:35  lr: 0.000273  min_lr: 0.000273  loss: 2.7552 (2.7552)  class_acc: 0.5938 (0.5938)  weight_decay: 0.0500 (0.0500)  time: 21.5616  data: 19.5654  max mem: 2905
Epoch: [253]  [200/625]  eta: 0:14:03  lr: 0.000269  min_lr: 0.000269  loss: 2.7445 (2.7441)  class_acc: 0.5820 (0.5851)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.0631 (3.2907)  time: 2.0001  data: 0.0007  max mem: 2905
Epoch: [253]  [400/625]  eta: 0:07:21  lr: 0.000265  min_lr: 0.000265  loss: 2.7319 (2.7394)  class_acc: 0.5781 (0.5864)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.4387 (inf)  time: 1.8362  data: 0.0007  max mem: 2905
Epoch: [253]  [600/625]  eta: 0:00:49  lr: 0.000262  min_lr: 0.000262  loss: 2.7257 (2.7411)  class_acc: 0.6055 (0.5868)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4727 (inf)  time: 2.0141  data: 0.0007  max mem: 2905
Epoch: [253]  [624/625]  eta: 0:00:01  lr: 0.000261  min_lr: 0.000261  loss: 2.7982 (2.7432)  class_acc: 0.5703 (0.5863)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2750 (inf)  time: 0.6853  data: 0.0014  max mem: 2905
Epoch: [253] Total time: 0:20:27 (1.9635 s / it)
Averaged stats: lr: 0.000261  min_lr: 0.000261  loss: 2.7982 (2.7398)  class_acc: 0.5703 (0.5866)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2750 (inf)
Test:  [ 0/50]  eta: 0:10:12  loss: 1.5349 (1.5349)  acc1: 63.2000 (63.2000)  acc5: 89.6000 (89.6000)  time: 12.2600  data: 12.2342  max mem: 2905
Test:  [10/50]  eta: 0:01:11  loss: 1.4989 (1.4906)  acc1: 67.2000 (67.0545)  acc5: 85.6000 (86.1091)  time: 1.7872  data: 1.7648  max mem: 2905
Test:  [20/50]  eta: 0:00:40  loss: 1.5942 (1.6474)  acc1: 62.4000 (62.8952)  acc5: 84.0000 (84.5714)  time: 0.8206  data: 0.8002  max mem: 2905
Test:  [30/50]  eta: 0:00:25  loss: 1.7724 (1.6507)  acc1: 60.0000 (62.8645)  acc5: 83.2000 (84.4387)  time: 0.9579  data: 0.9394  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 1.7413 (1.6688)  acc1: 61.6000 (62.8488)  acc5: 83.2000 (84.2342)  time: 0.9932  data: 0.9747  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 1.5952 (1.6775)  acc1: 61.6000 (62.8000)  acc5: 82.4000 (84.0960)  time: 0.7428  data: 0.7238  max mem: 2905
Test: Total time: 0:00:53 (1.0765 s / it)
* Acc@1 63.346 Acc@5 84.782 loss 1.643
Accuracy of the model on the 50000 test images: 63.3%
Max accuracy: 64.75%
Epoch: [254]  [  0/625]  eta: 3:44:14  lr: 0.000261  min_lr: 0.000261  loss: 2.7654 (2.7654)  class_acc: 0.5859 (0.5859)  weight_decay: 0.0500 (0.0500)  time: 21.5264  data: 19.3171  max mem: 2905
Epoch: [254]  [200/625]  eta: 0:14:53  lr: 0.000258  min_lr: 0.000258  loss: 2.7006 (2.7169)  class_acc: 0.5820 (0.5924)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.4051 (3.5633)  time: 2.1138  data: 0.3264  max mem: 2905
Epoch: [254]  [400/625]  eta: 0:07:36  lr: 0.000254  min_lr: 0.000254  loss: 2.6711 (2.7241)  class_acc: 0.5859 (0.5898)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5500 (3.4284)  time: 1.9595  data: 0.0126  max mem: 2905
Epoch: [254]  [600/625]  eta: 0:00:50  lr: 0.000251  min_lr: 0.000251  loss: 2.7435 (2.7343)  class_acc: 0.5820 (0.5884)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5560 (3.4639)  time: 2.1111  data: 0.0007  max mem: 2905
Epoch: [254]  [624/625]  eta: 0:00:01  lr: 0.000251  min_lr: 0.000251  loss: 2.7208 (2.7339)  class_acc: 0.5898 (0.5887)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7047 (3.4401)  time: 0.7457  data: 0.0072  max mem: 2905
Epoch: [254] Total time: 0:20:31 (1.9702 s / it)
Averaged stats: lr: 0.000251  min_lr: 0.000251  loss: 2.7208 (2.7364)  class_acc: 0.5898 (0.5874)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7047 (3.4401)
Test:  [ 0/50]  eta: 0:10:10  loss: 1.4909 (1.4909)  acc1: 64.0000 (64.0000)  acc5: 88.0000 (88.0000)  time: 12.2019  data: 12.1770  max mem: 2905
Test:  [10/50]  eta: 0:01:24  loss: 1.3238 (1.3467)  acc1: 69.6000 (70.0364)  acc5: 88.0000 (88.2182)  time: 2.1171  data: 2.0980  max mem: 2905
Test:  [20/50]  eta: 0:00:46  loss: 1.4950 (1.5084)  acc1: 66.4000 (66.5143)  acc5: 86.4000 (86.2476)  time: 1.0215  data: 1.0032  max mem: 2905
Test:  [30/50]  eta: 0:00:26  loss: 1.6170 (1.5169)  acc1: 64.0000 (65.9355)  acc5: 84.8000 (86.3226)  time: 0.8513  data: 0.8334  max mem: 2905
Test:  [40/50]  eta: 0:00:11  loss: 1.5418 (1.5407)  acc1: 62.4000 (65.4634)  acc5: 84.8000 (85.9122)  time: 0.7529  data: 0.7343  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 1.5919 (1.5532)  acc1: 63.2000 (64.9600)  acc5: 84.8000 (85.6960)  time: 0.6818  data: 0.6626  max mem: 2905
Test: Total time: 0:00:51 (1.0252 s / it)
* Acc@1 65.522 Acc@5 86.162 loss 1.520
Accuracy of the model on the 50000 test images: 65.5%
Max accuracy: 65.52%
Epoch: [255]  [  0/625]  eta: 3:26:30  lr: 0.000250  min_lr: 0.000250  loss: 2.8605 (2.8605)  class_acc: 0.5469 (0.5469)  weight_decay: 0.0500 (0.0500)  time: 19.8255  data: 19.4243  max mem: 2905
Epoch: [255]  [200/625]  eta: 0:14:09  lr: 0.000247  min_lr: 0.000247  loss: 2.7077 (2.7239)  class_acc: 0.5977 (0.5927)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.8561 (3.1735)  time: 1.9528  data: 0.3191  max mem: 2905
Epoch: [255]  [400/625]  eta: 0:07:20  lr: 0.000244  min_lr: 0.000244  loss: 2.6997 (2.7316)  class_acc: 0.6094 (0.5896)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6714 (3.2675)  time: 1.9264  data: 0.0013  max mem: 2905
Epoch: [255]  [600/625]  eta: 0:00:48  lr: 0.000240  min_lr: 0.000240  loss: 2.7242 (2.7341)  class_acc: 0.5938 (0.5891)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.7159 (3.2027)  time: 1.7563  data: 0.0066  max mem: 2905
Epoch: [255]  [624/625]  eta: 0:00:01  lr: 0.000240  min_lr: 0.000240  loss: 2.7479 (2.7336)  class_acc: 0.5820 (0.5893)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.4637 (3.1991)  time: 1.1358  data: 0.0260  max mem: 2905
Epoch: [255] Total time: 0:20:04 (1.9264 s / it)
Averaged stats: lr: 0.000240  min_lr: 0.000240  loss: 2.7479 (2.7322)  class_acc: 0.5820 (0.5884)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.4637 (3.1991)
Test:  [ 0/50]  eta: 0:08:20  loss: 1.5699 (1.5699)  acc1: 62.4000 (62.4000)  acc5: 88.0000 (88.0000)  time: 10.0079  data: 9.9798  max mem: 2905
Test:  [10/50]  eta: 0:01:09  loss: 1.4135 (1.4245)  acc1: 67.2000 (67.9273)  acc5: 87.2000 (87.3455)  time: 1.7276  data: 1.7048  max mem: 2905
Test:  [20/50]  eta: 0:00:38  loss: 1.5388 (1.5779)  acc1: 65.6000 (64.9905)  acc5: 85.6000 (85.1810)  time: 0.8387  data: 0.8168  max mem: 2905
Test:  [30/50]  eta: 0:00:22  loss: 1.7482 (1.5958)  acc1: 61.6000 (64.3871)  acc5: 85.6000 (85.2645)  time: 0.8141  data: 0.7946  max mem: 2905
Test:  [40/50]  eta: 0:00:10  loss: 1.6335 (1.6109)  acc1: 61.6000 (64.1756)  acc5: 85.6000 (84.8585)  time: 0.8951  data: 0.8764  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 1.6059 (1.6096)  acc1: 64.8000 (64.3200)  acc5: 85.6000 (84.9760)  time: 0.6413  data: 0.6216  max mem: 2905
Test: Total time: 0:00:48 (0.9653 s / it)
* Acc@1 65.094 Acc@5 85.712 loss 1.567
Accuracy of the model on the 50000 test images: 65.1%
Max accuracy: 65.52%
Epoch: [256]  [  0/625]  eta: 3:33:46  lr: 0.000240  min_lr: 0.000240  loss: 2.8370 (2.8370)  class_acc: 0.5703 (0.5703)  weight_decay: 0.0500 (0.0500)  time: 20.5227  data: 18.6952  max mem: 2905
Epoch: [256]  [200/625]  eta: 0:13:50  lr: 0.000236  min_lr: 0.000236  loss: 2.6545 (2.7142)  class_acc: 0.5859 (0.5924)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5862 (3.1751)  time: 1.6284  data: 0.0050  max mem: 2905
Epoch: [256]  [400/625]  eta: 0:07:14  lr: 0.000233  min_lr: 0.000233  loss: 2.7374 (2.7253)  class_acc: 0.5859 (0.5902)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7179 (3.0621)  time: 1.9977  data: 0.0008  max mem: 2905
Epoch: [256]  [600/625]  eta: 0:00:47  lr: 0.000230  min_lr: 0.000230  loss: 2.7958 (2.7376)  class_acc: 0.5703 (0.5874)  weight_decay: 0.0500 (0.0500)  grad_norm: 4.5650 (3.5334)  time: 1.8449  data: 0.0008  max mem: 2905
Epoch: [256]  [624/625]  eta: 0:00:01  lr: 0.000229  min_lr: 0.000229  loss: 2.7231 (2.7376)  class_acc: 0.5781 (0.5872)  weight_decay: 0.0500 (0.0500)  grad_norm: 4.7181 (3.5818)  time: 0.7537  data: 0.0015  max mem: 2905
Epoch: [256] Total time: 0:19:44 (1.8956 s / it)
Averaged stats: lr: 0.000229  min_lr: 0.000229  loss: 2.7231 (2.7325)  class_acc: 0.5781 (0.5886)  weight_decay: 0.0500 (0.0500)  grad_norm: 4.7181 (3.5818)
Test:  [ 0/50]  eta: 0:10:22  loss: 1.4078 (1.4078)  acc1: 64.0000 (64.0000)  acc5: 91.2000 (91.2000)  time: 12.4423  data: 12.4060  max mem: 2905
Test:  [10/50]  eta: 0:01:22  loss: 1.4078 (1.3778)  acc1: 68.8000 (69.7455)  acc5: 88.8000 (88.1455)  time: 2.0697  data: 2.0498  max mem: 2905
Test:  [20/50]  eta: 0:00:45  loss: 1.5660 (1.5334)  acc1: 64.8000 (65.7905)  acc5: 86.4000 (86.4381)  time: 0.9580  data: 0.9385  max mem: 2905
Test:  [30/50]  eta: 0:00:24  loss: 1.6332 (1.5371)  acc1: 63.2000 (65.4194)  acc5: 85.6000 (86.1677)  time: 0.7938  data: 0.7741  max mem: 2905
Test:  [40/50]  eta: 0:00:10  loss: 1.5259 (1.5596)  acc1: 64.0000 (65.0732)  acc5: 85.6000 (85.7561)  time: 0.6158  data: 0.5961  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 1.5176 (1.5672)  acc1: 64.0000 (64.8960)  acc5: 84.8000 (85.5520)  time: 0.5330  data: 0.5138  max mem: 2905
Test: Total time: 0:00:46 (0.9385 s / it)
* Acc@1 65.690 Acc@5 86.250 loss 1.534
Accuracy of the model on the 50000 test images: 65.7%
Max accuracy: 65.69%
Epoch: [257]  [  0/625]  eta: 3:45:18  lr: 0.000229  min_lr: 0.000229  loss: 2.5982 (2.5982)  class_acc: 0.6133 (0.6133)  weight_decay: 0.0500 (0.0500)  time: 21.6303  data: 16.3646  max mem: 2905
Epoch: [257]  [200/625]  eta: 0:14:25  lr: 0.000226  min_lr: 0.000226  loss: 2.7106 (2.7174)  class_acc: 0.5820 (0.5894)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.2095 (3.2202)  time: 1.7910  data: 0.0011  max mem: 2905
Epoch: [257]  [400/625]  eta: 0:07:27  lr: 0.000223  min_lr: 0.000223  loss: 2.7566 (2.7293)  class_acc: 0.5742 (0.5893)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.3751 (3.3419)  time: 1.9253  data: 0.0008  max mem: 2905
Epoch: [257]  [600/625]  eta: 0:00:49  lr: 0.000219  min_lr: 0.000219  loss: 2.6901 (2.7289)  class_acc: 0.5859 (0.5891)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6949 (3.3159)  time: 2.0633  data: 0.0008  max mem: 2905
Epoch: [257]  [624/625]  eta: 0:00:01  lr: 0.000219  min_lr: 0.000219  loss: 2.7255 (2.7292)  class_acc: 0.5742 (0.5889)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3787 (3.2976)  time: 0.8807  data: 0.0018  max mem: 2905
Epoch: [257] Total time: 0:20:15 (1.9455 s / it)
Averaged stats: lr: 0.000219  min_lr: 0.000219  loss: 2.7255 (2.7289)  class_acc: 0.5742 (0.5894)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3787 (3.2976)
Test:  [ 0/50]  eta: 0:11:05  loss: 1.5713 (1.5713)  acc1: 64.0000 (64.0000)  acc5: 87.2000 (87.2000)  time: 13.3129  data: 13.2860  max mem: 2905
Test:  [10/50]  eta: 0:01:27  loss: 1.3897 (1.4096)  acc1: 69.6000 (69.9636)  acc5: 88.8000 (88.0727)  time: 2.1845  data: 2.1655  max mem: 2905
Test:  [20/50]  eta: 0:00:47  loss: 1.5689 (1.5329)  acc1: 66.4000 (66.7048)  acc5: 87.2000 (87.0095)  time: 1.0032  data: 0.9843  max mem: 2905
Test:  [30/50]  eta: 0:00:26  loss: 1.6281 (1.5558)  acc1: 64.8000 (66.3484)  acc5: 86.4000 (86.3742)  time: 0.8768  data: 0.8575  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 1.5405 (1.5741)  acc1: 64.8000 (65.6390)  acc5: 85.6000 (85.9512)  time: 0.8397  data: 0.8204  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 1.5772 (1.5809)  acc1: 64.0000 (65.5520)  acc5: 85.6000 (85.8080)  time: 0.7819  data: 0.7610  max mem: 2905
Test: Total time: 0:00:54 (1.0885 s / it)
* Acc@1 65.784 Acc@5 86.180 loss 1.553
Accuracy of the model on the 50000 test images: 65.8%
Max accuracy: 65.78%
Epoch: [258]  [  0/625]  eta: 4:14:33  lr: 0.000219  min_lr: 0.000219  loss: 2.7456 (2.7456)  class_acc: 0.5703 (0.5703)  weight_decay: 0.0500 (0.0500)  time: 24.4371  data: 19.5015  max mem: 2905
Epoch: [258]  [200/625]  eta: 0:14:19  lr: 0.000216  min_lr: 0.000216  loss: 2.6921 (2.7116)  class_acc: 0.5938 (0.5936)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.3964 (3.2895)  time: 1.8386  data: 0.0007  max mem: 2905
Epoch: [258]  [400/625]  eta: 0:07:22  lr: 0.000212  min_lr: 0.000212  loss: 2.6857 (2.7084)  class_acc: 0.5938 (0.5940)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.9483 (3.4547)  time: 1.7934  data: 0.0007  max mem: 2905
Epoch: [258]  [600/625]  eta: 0:00:48  lr: 0.000209  min_lr: 0.000209  loss: 2.6951 (2.7124)  class_acc: 0.5938 (0.5930)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6143 (3.3969)  time: 1.9270  data: 0.0007  max mem: 2905
Epoch: [258]  [624/625]  eta: 0:00:01  lr: 0.000209  min_lr: 0.000209  loss: 2.7552 (2.7144)  class_acc: 0.5820 (0.5926)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7592 (3.3967)  time: 0.7888  data: 0.0016  max mem: 2905
Epoch: [258] Total time: 0:20:11 (1.9381 s / it)
Averaged stats: lr: 0.000209  min_lr: 0.000209  loss: 2.7552 (2.7240)  class_acc: 0.5820 (0.5906)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7592 (3.3967)
Test:  [ 0/50]  eta: 0:09:24  loss: 1.5589 (1.5589)  acc1: 64.0000 (64.0000)  acc5: 84.0000 (84.0000)  time: 11.2899  data: 11.2559  max mem: 2905
Test:  [10/50]  eta: 0:01:08  loss: 1.4893 (1.4482)  acc1: 65.6000 (67.0545)  acc5: 87.2000 (87.2000)  time: 1.7222  data: 1.7018  max mem: 2905
Test:  [20/50]  eta: 0:00:36  loss: 1.5755 (1.6072)  acc1: 63.2000 (64.2286)  acc5: 86.4000 (85.8286)  time: 0.7101  data: 0.6907  max mem: 2905
Test:  [30/50]  eta: 0:00:23  loss: 1.7574 (1.6328)  acc1: 60.8000 (63.5097)  acc5: 85.6000 (85.4710)  time: 0.8587  data: 0.8384  max mem: 2905
Test:  [40/50]  eta: 0:00:10  loss: 1.7542 (1.6624)  acc1: 60.8000 (63.0439)  acc5: 84.0000 (84.7610)  time: 0.8250  data: 0.8035  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 1.6714 (1.6670)  acc1: 60.8000 (62.8000)  acc5: 83.2000 (84.6400)  time: 0.4191  data: 0.3987  max mem: 2905
Test: Total time: 0:00:44 (0.8928 s / it)
* Acc@1 63.704 Acc@5 84.766 loss 1.631
Accuracy of the model on the 50000 test images: 63.7%
Max accuracy: 65.78%
Epoch: [259]  [  0/625]  eta: 3:26:00  lr: 0.000209  min_lr: 0.000209  loss: 2.9102 (2.9102)  class_acc: 0.5625 (0.5625)  weight_decay: 0.0500 (0.0500)  time: 19.7768  data: 18.1604  max mem: 2905
Epoch: [259]  [200/625]  eta: 0:13:43  lr: 0.000206  min_lr: 0.000206  loss: 2.7059 (2.7221)  class_acc: 0.5859 (0.5896)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.9464 (3.4944)  time: 1.8997  data: 0.0118  max mem: 2905
Epoch: [259]  [400/625]  eta: 0:07:13  lr: 0.000203  min_lr: 0.000203  loss: 2.6808 (2.7276)  class_acc: 0.5977 (0.5881)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7679 (3.5310)  time: 1.9214  data: 0.0136  max mem: 2905
Epoch: [259]  [600/625]  eta: 0:00:48  lr: 0.000199  min_lr: 0.000199  loss: 2.7244 (2.7279)  class_acc: 0.5742 (0.5881)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.8087 (3.4387)  time: 2.1257  data: 0.0089  max mem: 2905
Epoch: [259]  [624/625]  eta: 0:00:01  lr: 0.000199  min_lr: 0.000199  loss: 2.7092 (2.7275)  class_acc: 0.5859 (0.5883)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6717 (3.4193)  time: 0.8008  data: 0.0017  max mem: 2905
Epoch: [259] Total time: 0:20:03 (1.9258 s / it)
Averaged stats: lr: 0.000199  min_lr: 0.000199  loss: 2.7092 (2.7235)  class_acc: 0.5859 (0.5904)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6717 (3.4193)
Test:  [ 0/50]  eta: 0:10:44  loss: 1.3960 (1.3960)  acc1: 69.6000 (69.6000)  acc5: 87.2000 (87.2000)  time: 12.8839  data: 12.8585  max mem: 2905
Test:  [10/50]  eta: 0:01:30  loss: 1.3465 (1.3265)  acc1: 69.6000 (71.5636)  acc5: 88.0000 (88.2182)  time: 2.2662  data: 2.2435  max mem: 2905
Test:  [20/50]  eta: 0:00:54  loss: 1.4350 (1.4891)  acc1: 68.0000 (67.7333)  acc5: 87.2000 (86.7810)  time: 1.2611  data: 1.2393  max mem: 2905
Test:  [30/50]  eta: 0:00:32  loss: 1.6341 (1.4996)  acc1: 63.2000 (66.6323)  acc5: 85.6000 (86.5032)  time: 1.2560  data: 1.2345  max mem: 2905
Test:  [40/50]  eta: 0:00:13  loss: 1.4941 (1.5167)  acc1: 62.4000 (65.9902)  acc5: 85.6000 (86.1268)  time: 0.8591  data: 0.8385  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 1.4805 (1.5255)  acc1: 62.4000 (65.5520)  acc5: 85.6000 (85.9360)  time: 0.7805  data: 0.7609  max mem: 2905
Test: Total time: 0:00:56 (1.1286 s / it)
* Acc@1 66.380 Acc@5 86.542 loss 1.490
Accuracy of the model on the 50000 test images: 66.4%
Max accuracy: 66.38%
Epoch: [260]  [  0/625]  eta: 3:19:49  lr: 0.000199  min_lr: 0.000199  loss: 2.7660 (2.7660)  class_acc: 0.5391 (0.5391)  weight_decay: 0.0500 (0.0500)  time: 19.1827  data: 17.9289  max mem: 2905
Epoch: [260]  [200/625]  eta: 0:14:19  lr: 0.000196  min_lr: 0.000196  loss: 2.7087 (2.7025)  class_acc: 0.5898 (0.5976)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.3598 (3.7958)  time: 1.9344  data: 0.0006  max mem: 2905
Epoch: [260]  [400/625]  eta: 0:07:22  lr: 0.000193  min_lr: 0.000193  loss: 2.7273 (2.7128)  class_acc: 0.5898 (0.5934)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5857 (inf)  time: 1.8016  data: 0.0007  max mem: 2905
Epoch: [260]  [600/625]  eta: 0:00:49  lr: 0.000190  min_lr: 0.000190  loss: 2.7195 (2.7200)  class_acc: 0.5859 (0.5914)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.6396 (inf)  time: 2.0866  data: 0.0006  max mem: 2905
Epoch: [260]  [624/625]  eta: 0:00:01  lr: 0.000189  min_lr: 0.000189  loss: 2.7451 (2.7209)  class_acc: 0.5859 (0.5914)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.5698 (inf)  time: 0.5684  data: 0.0017  max mem: 2905
Epoch: [260] Total time: 0:20:25 (1.9616 s / it)
Averaged stats: lr: 0.000189  min_lr: 0.000189  loss: 2.7451 (2.7205)  class_acc: 0.5859 (0.5916)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.5698 (inf)
Test:  [ 0/50]  eta: 0:10:06  loss: 1.4237 (1.4237)  acc1: 69.6000 (69.6000)  acc5: 88.8000 (88.8000)  time: 12.1305  data: 12.0700  max mem: 2905
Test:  [10/50]  eta: 0:01:22  loss: 1.3327 (1.3290)  acc1: 71.2000 (70.7636)  acc5: 90.4000 (89.8182)  time: 2.0737  data: 2.0507  max mem: 2905
Test:  [20/50]  eta: 0:00:49  loss: 1.4568 (1.4776)  acc1: 66.4000 (66.8571)  acc5: 88.8000 (88.2286)  time: 1.1268  data: 1.1078  max mem: 2905
Test:  [30/50]  eta: 0:00:29  loss: 1.5571 (1.5026)  acc1: 64.8000 (66.4000)  acc5: 86.4000 (87.4581)  time: 1.1486  data: 1.1300  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 1.5411 (1.5282)  acc1: 64.8000 (65.7951)  acc5: 85.6000 (86.8488)  time: 0.9195  data: 0.9002  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 1.5188 (1.5343)  acc1: 64.8000 (65.6320)  acc5: 85.6000 (86.8160)  time: 0.8881  data: 0.8682  max mem: 2905
Test: Total time: 0:00:53 (1.0711 s / it)
* Acc@1 66.532 Acc@5 86.990 loss 1.497
Accuracy of the model on the 50000 test images: 66.5%
Max accuracy: 66.53%
Epoch: [261]  [  0/625]  eta: 4:28:55  lr: 0.000189  min_lr: 0.000189  loss: 2.5993 (2.5993)  class_acc: 0.6172 (0.6172)  weight_decay: 0.0500 (0.0500)  time: 25.8162  data: 18.7189  max mem: 2905
Epoch: [261]  [200/625]  eta: 0:14:47  lr: 0.000186  min_lr: 0.000186  loss: 2.7165 (2.7027)  class_acc: 0.5742 (0.5928)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.2815 (3.5401)  time: 1.8274  data: 0.0007  max mem: 2905
Epoch: [261]  [400/625]  eta: 0:07:43  lr: 0.000183  min_lr: 0.000183  loss: 2.6623 (2.7087)  class_acc: 0.5938 (0.5927)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.9501 (3.4142)  time: 2.0727  data: 0.0006  max mem: 2905
Epoch: [261]  [600/625]  eta: 0:00:51  lr: 0.000180  min_lr: 0.000180  loss: 2.7268 (2.7143)  class_acc: 0.5859 (0.5919)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.2833 (3.3762)  time: 2.0241  data: 0.0007  max mem: 2905
Epoch: [261]  [624/625]  eta: 0:00:01  lr: 0.000180  min_lr: 0.000180  loss: 2.7475 (2.7152)  class_acc: 0.5859 (0.5917)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7709 (3.3768)  time: 0.8703  data: 0.0013  max mem: 2905
Epoch: [261] Total time: 0:20:50 (2.0001 s / it)
Averaged stats: lr: 0.000180  min_lr: 0.000180  loss: 2.7475 (2.7206)  class_acc: 0.5859 (0.5913)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7709 (3.3768)
Test:  [ 0/50]  eta: 0:10:15  loss: 1.4869 (1.4869)  acc1: 63.2000 (63.2000)  acc5: 86.4000 (86.4000)  time: 12.3170  data: 12.2876  max mem: 2905
Test:  [10/50]  eta: 0:01:28  loss: 1.3026 (1.3153)  acc1: 71.2000 (70.5455)  acc5: 88.8000 (88.2909)  time: 2.2218  data: 2.2016  max mem: 2905
Test:  [20/50]  eta: 0:00:54  loss: 1.4055 (1.4615)  acc1: 67.2000 (67.3905)  acc5: 87.2000 (86.8952)  time: 1.2754  data: 1.2548  max mem: 2905
Test:  [30/50]  eta: 0:00:31  loss: 1.5319 (1.4809)  acc1: 63.2000 (66.5032)  acc5: 87.2000 (86.8387)  time: 1.2323  data: 1.2110  max mem: 2905
Test:  [40/50]  eta: 0:00:13  loss: 1.4787 (1.5068)  acc1: 63.2000 (66.1073)  acc5: 84.8000 (86.3220)  time: 0.8046  data: 0.7844  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 1.5062 (1.5128)  acc1: 64.8000 (65.7920)  acc5: 84.8000 (86.2240)  time: 0.7535  data: 0.7338  max mem: 2905
Test: Total time: 0:00:54 (1.0987 s / it)
* Acc@1 66.388 Acc@5 86.686 loss 1.484
Accuracy of the model on the 50000 test images: 66.4%
Max accuracy: 66.53%
Epoch: [262]  [  0/625]  eta: 3:32:44  lr: 0.000180  min_lr: 0.000180  loss: 2.6006 (2.6006)  class_acc: 0.5859 (0.5859)  weight_decay: 0.0500 (0.0500)  time: 20.4236  data: 18.8559  max mem: 2905
Epoch: [262]  [200/625]  eta: 0:14:25  lr: 0.000177  min_lr: 0.000177  loss: 2.6758 (2.7199)  class_acc: 0.6094 (0.5920)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.0267 (3.4221)  time: 1.9948  data: 0.0271  max mem: 2905
Epoch: [262]  [400/625]  eta: 0:07:32  lr: 0.000174  min_lr: 0.000174  loss: 2.7035 (2.7177)  class_acc: 0.5977 (0.5923)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5609 (3.1984)  time: 2.0504  data: 0.0015  max mem: 2905
Epoch: [262]  [600/625]  eta: 0:00:50  lr: 0.000171  min_lr: 0.000171  loss: 2.6726 (2.7195)  class_acc: 0.5898 (0.5923)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.9992 (3.2388)  time: 2.0665  data: 0.0009  max mem: 2905
Epoch: [262]  [624/625]  eta: 0:00:01  lr: 0.000171  min_lr: 0.000171  loss: 2.6545 (2.7181)  class_acc: 0.5938 (0.5923)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.9992 (3.2367)  time: 0.8305  data: 0.0014  max mem: 2905
Epoch: [262] Total time: 0:20:28 (1.9650 s / it)
Averaged stats: lr: 0.000171  min_lr: 0.000171  loss: 2.6545 (2.7193)  class_acc: 0.5938 (0.5919)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.9992 (3.2367)
Test:  [ 0/50]  eta: 0:10:59  loss: 1.3657 (1.3657)  acc1: 64.0000 (64.0000)  acc5: 89.6000 (89.6000)  time: 13.1889  data: 13.1503  max mem: 2905
Test:  [10/50]  eta: 0:01:26  loss: 1.3222 (1.3373)  acc1: 70.4000 (70.9818)  acc5: 89.6000 (88.9455)  time: 2.1616  data: 2.1409  max mem: 2905
Test:  [20/50]  eta: 0:00:50  loss: 1.4960 (1.4961)  acc1: 65.6000 (67.0476)  acc5: 87.2000 (87.1619)  time: 1.1172  data: 1.0983  max mem: 2905
Test:  [30/50]  eta: 0:00:29  loss: 1.5735 (1.5061)  acc1: 63.2000 (66.6581)  acc5: 86.4000 (87.0194)  time: 1.1156  data: 1.0956  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 1.4596 (1.5319)  acc1: 64.0000 (66.1073)  acc5: 86.4000 (86.4000)  time: 0.8460  data: 0.8255  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 1.5491 (1.5417)  acc1: 64.0000 (65.6480)  acc5: 86.4000 (86.1920)  time: 0.8027  data: 0.7828  max mem: 2905
Test: Total time: 0:00:53 (1.0635 s / it)
* Acc@1 66.068 Acc@5 86.440 loss 1.514
Accuracy of the model on the 50000 test images: 66.1%
Max accuracy: 66.53%
Epoch: [263]  [  0/625]  eta: 4:08:20  lr: 0.000171  min_lr: 0.000171  loss: 2.5159 (2.5159)  class_acc: 0.6367 (0.6367)  weight_decay: 0.0500 (0.0500)  time: 23.8410  data: 18.6652  max mem: 2905
Epoch: [263]  [200/625]  eta: 0:14:31  lr: 0.000168  min_lr: 0.000168  loss: 2.7019 (2.7243)  class_acc: 0.5938 (0.5909)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.6454 (3.5457)  time: 1.9884  data: 0.8526  max mem: 2905
Epoch: [263]  [400/625]  eta: 0:07:26  lr: 0.000165  min_lr: 0.000165  loss: 2.6986 (2.7227)  class_acc: 0.5977 (0.5907)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.0232 (3.3814)  time: 2.0851  data: 0.0008  max mem: 2905
Epoch: [263]  [600/625]  eta: 0:00:49  lr: 0.000162  min_lr: 0.000162  loss: 2.6886 (2.7212)  class_acc: 0.5859 (0.5907)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7652 (3.3005)  time: 1.8322  data: 0.0061  max mem: 2905
Epoch: [263]  [624/625]  eta: 0:00:01  lr: 0.000162  min_lr: 0.000162  loss: 2.7270 (2.7223)  class_acc: 0.5781 (0.5904)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.8724 (3.2828)  time: 0.7270  data: 0.0014  max mem: 2905
Epoch: [263] Total time: 0:20:11 (1.9385 s / it)
Averaged stats: lr: 0.000162  min_lr: 0.000162  loss: 2.7270 (2.7147)  class_acc: 0.5781 (0.5925)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.8724 (3.2828)
Test:  [ 0/50]  eta: 0:09:35  loss: 1.4544 (1.4544)  acc1: 65.6000 (65.6000)  acc5: 88.8000 (88.8000)  time: 11.5030  data: 11.4767  max mem: 2905
Test:  [10/50]  eta: 0:01:11  loss: 1.3297 (1.3302)  acc1: 68.8000 (70.6182)  acc5: 89.6000 (89.5273)  time: 1.7841  data: 1.7630  max mem: 2905
Test:  [20/50]  eta: 0:00:40  loss: 1.4719 (1.4732)  acc1: 66.4000 (67.0857)  acc5: 88.0000 (87.8095)  time: 0.8488  data: 0.8290  max mem: 2905
Test:  [30/50]  eta: 0:00:24  loss: 1.5536 (1.4999)  acc1: 64.0000 (66.5806)  acc5: 86.4000 (87.0452)  time: 0.9526  data: 0.9320  max mem: 2905
Test:  [40/50]  eta: 0:00:10  loss: 1.4940 (1.5238)  acc1: 64.8000 (65.9122)  acc5: 85.6000 (86.4781)  time: 0.7851  data: 0.7650  max mem: 2905
Test:  [49/50]  eta: 0:00:00  loss: 1.5341 (1.5228)  acc1: 65.6000 (65.6800)  acc5: 85.6000 (86.4640)  time: 0.4189  data: 0.4007  max mem: 2905
Test: Total time: 0:00:47 (0.9438 s / it)
* Acc@1 66.454 Acc@5 86.734 loss 1.496
Accuracy of the model on the 50000 test images: 66.5%
Max accuracy: 66.53%
Epoch: [264]  [  0/625]  eta: 3:24:32  lr: 0.000162  min_lr: 0.000162  loss: 2.7634 (2.7634)  class_acc: 0.5898 (0.5898)  weight_decay: 0.0500 (0.0500)  time: 19.6367  data: 17.2441  max mem: 2905
Epoch: [264]  [200/625]  eta: 0:13:58  lr: 0.000159  min_lr: 0.000159  loss: 2.7183 (2.7173)  class_acc: 0.5820 (0.5933)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4330 (3.4661)  time: 1.8096  data: 0.6035  max mem: 2905
Epoch: [264]  [400/625]  eta: 0:07:17  lr: 0.000156  min_lr: 0.000156  loss: 2.7310 (2.7119)  class_acc: 0.5859 (0.5933)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.8214 (3.4802)  time: 1.9498  data: 0.0954  max mem: 2905
Epoch: [264]  [600/625]  eta: 0:00:48  lr: 0.000154  min_lr: 0.000154  loss: 2.6837 (2.7124)  class_acc: 0.5977 (0.5935)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.8499 (3.5127)  time: 2.0400  data: 0.1734  max mem: 2905
Epoch: [264]  [624/625]  eta: 0:00:01  lr: 0.000153  min_lr: 0.000153  loss: 2.6487 (2.7113)  class_acc: 0.6016 (0.5938)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7755 (3.4780)  time: 0.7590  data: 0.0324  max mem: 2905
Epoch: [264] Total time: 0:19:40 (1.8882 s / it)
Averaged stats: lr: 0.000153  min_lr: 0.000153  loss: 2.6487 (2.7131)  class_acc: 0.6016 (0.5927)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7755 (3.4780)
Test:  [ 0/50]  eta: 0:10:46  loss: 1.4714 (1.4714)  acc1: 65.6000 (65.6000)  acc5: 87.2000 (87.2000)  time: 12.9237  data: 12.8982  max mem: 2905
Test:  [10/50]  eta: 0:01:17  loss: 1.3041 (1.2939)  acc1: 69.6000 (71.0545)  acc5: 89.6000 (88.6546)  time: 1.9316  data: 1.9128  max mem: 2905
Test:  [20/50]  eta: 0:00:42  loss: 1.4484 (1.4556)  acc1: 68.0000 (67.6952)  acc5: 87.2000 (86.9714)  time: 0.8514  data: 0.8328  max mem: 2905
Test:  [30/50]  eta: 0:00:26  loss: 1.5503 (1.4833)  acc1: 64.8000 (66.9677)  acc5: 85.6000 (86.7613)  time: 0.9835  data: 0.9645  max mem: 2905
Test:  [40/50]  eta: 0:00:12  loss: 1.5289 (1.5031)  acc1: 64.0000 (66.4585)  acc5: 85.6000 (86.2439)  time: 0.9746  data: 0.9549  max mem: 2905
Test:  [49/50]  eta: 0:00:01  loss: 1.4980 (1.5058)  acc1: 65.6000 (66.4160)  acc5: 85.6000 (86.1120)  time: 0.6406  data: 0.6217  max mem: 2905
Test: Total time: 0:00:53 (1.0784 s / it)
* Acc@1 66.778 Acc@5 86.864 loss 1.475
Accuracy of the model on the 50000 test images: 66.8%
Max accuracy: 66.78%
Epoch: [265]  [  0/625]  eta: 3:31:23  lr: 0.000153  min_lr: 0.000153  loss: 2.7851 (2.7851)  class_acc: 0.5859 (0.5859)  weight_decay: 0.0500 (0.0500)  time: 20.2935  data: 19.8328  max mem: 2905
WARNING:__main__:
*****************************************
Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. 
*****************************************
| distributed init (rank 0): env://, gpu 0
| distributed init (rank 5): env://, gpu 5
| distributed init (rank 6): env://, gpu 6
| distributed init (rank 3): env://, gpu 3
| distributed init (rank 7): env://, gpu 7
| distributed init (rank 1): env://, gpu 1
| distributed init (rank 4): env://, gpu 4
| distributed init (rank 2): env://, gpu 2
Namespace(aa='rand-m9-mstd0.5-inc1', auto_resume=True, batch_size=256, clip_grad=None, color_jitter=0.4, crop_pct=None, cutmix=0.0, cutmix_minmax=None, data_path='/data/benchmarks/ILSVRC2012_LMDB', data_set='IMNET_LMDB', device='cuda', disable_eval=False, dist_backend='nccl', dist_eval=True, dist_on_itp=False, dist_url='env://', distributed=True, drop_path=0.2, enable_wandb=False, epochs=300, eval=False, eval_data_path=None, finetune='', gpu=0, head_init_scale=1.0, imagenet_default_mean_and_std=True, input_size=224, layer_decay=1.0, layer_scale_init_value=1e-06, local_rank=-1, log_dir=None, lr=0.004, min_lr=1e-06, mixup=0.0, mixup_mode='batch', mixup_prob=1.0, mixup_switch_prob=0.5, model='convnext_tiny', model_ema=False, model_ema_decay=0.9999, model_ema_eval=False, model_ema_force_cpu=False, model_key='model|module', model_prefix='', momentum=0.9, nb_classes=1000, num_workers=10, opt='adamw', opt_betas=None, opt_eps=1e-08, output_dir='./checkpoint', pin_mem=True, project='convnext', rank=0, recount=1, remode='pixel', reprob=0.25, resplit=False, resume='', save_ckpt=True, save_ckpt_freq=1, save_ckpt_num=3, seed=0, smoothing=0.1, start_epoch=0, train_interpolation='bicubic', update_freq=2, use_amp=True, wandb_ckpt=False, warmup_epochs=20, warmup_steps=-1, weight_decay=0.05, weight_decay_end=None, world_size=8)
Transform = 
RandomResizedCropAndInterpolation(size=(224, 224), scale=(0.08, 1.0), ratio=(0.75, 1.3333), interpolation=bicubic)
RandomHorizontalFlip(p=0.5)
RandAugment(n=2, ops=
	AugmentOp(name=AutoContrast, p=0.5, m=9, mstd=0.5)
	AugmentOp(name=Equalize, p=0.5, m=9, mstd=0.5)
	AugmentOp(name=Invert, p=0.5, m=9, mstd=0.5)
	AugmentOp(name=Rotate, p=0.5, m=9, mstd=0.5)
	AugmentOp(name=PosterizeIncreasing, p=0.5, m=9, mstd=0.5)
	AugmentOp(name=SolarizeIncreasing, p=0.5, m=9, mstd=0.5)
	AugmentOp(name=SolarizeAdd, p=0.5, m=9, mstd=0.5)
	AugmentOp(name=ColorIncreasing, p=0.5, m=9, mstd=0.5)
	AugmentOp(name=ContrastIncreasing, p=0.5, m=9, mstd=0.5)
	AugmentOp(name=BrightnessIncreasing, p=0.5, m=9, mstd=0.5)
	AugmentOp(name=SharpnessIncreasing, p=0.5, m=9, mstd=0.5)
	AugmentOp(name=ShearX, p=0.5, m=9, mstd=0.5)
	AugmentOp(name=ShearY, p=0.5, m=9, mstd=0.5)
	AugmentOp(name=TranslateXRel, p=0.5, m=9, mstd=0.5)
	AugmentOp(name=TranslateYRel, p=0.5, m=9, mstd=0.5))
ToTensor()
Normalize(mean=tensor([0.4850, 0.4560, 0.4060]), std=tensor([0.2290, 0.2240, 0.2250]))
RandomErasing(p=0.25, mode=pixel, count=(1, 1))
---------------------------
reading from datapath /data/benchmarks/ILSVRC2012_LMDB
Number of the class = 1000
Transform = 
Resize(size=256, interpolation=bicubic, max_size=None, antialias=None)
CenterCrop(size=(224, 224))
ToTensor()
Normalize(mean=(0.485, 0.456, 0.406), std=(0.229, 0.224, 0.225))
---------------------------
reading from datapath /data/benchmarks/ILSVRC2012_LMDB
Number of the class = 1000
Sampler_train = <torch.utils.data.distributed.DistributedSampler object at 0x7f763c0a9d60>
Model = MobileNetV3_Small(
  (conv1): Conv2d(3, 16, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
  (bn1): BatchNorm2d(16, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
  (hs1): Hardswish()
  (bneck): Sequential(
    (0): Block(
      (conv1): Conv2d(16, 16, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn1): BatchNorm2d(16, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (act1): ReLU(inplace=True)
      (conv2): Conv2d(16, 16, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), groups=16, bias=False)
      (bn2): BatchNorm2d(16, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (act2): ReLU(inplace=True)
      (se): SeModule(
        (se): Sequential(
          (0): AdaptiveAvgPool2d(output_size=1)
          (1): Conv2d(16, 8, kernel_size=(1, 1), stride=(1, 1), bias=False)
          (2): BatchNorm2d(8, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
          (3): ReLU(inplace=True)
          (4): Conv2d(8, 16, kernel_size=(1, 1), stride=(1, 1), bias=False)
          (5): Hardsigmoid()
        )
      )
      (conv3): Conv2d(16, 16, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn3): BatchNorm2d(16, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (act3): ReLU(inplace=True)
      (skip): Sequential(
        (0): Conv2d(16, 16, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), groups=16, bias=False)
        (1): BatchNorm2d(16, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      )
    )
    (1): Block(
      (conv1): Conv2d(16, 72, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn1): BatchNorm2d(72, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (act1): ReLU(inplace=True)
      (conv2): Conv2d(72, 72, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), groups=72, bias=False)
      (bn2): BatchNorm2d(72, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (act2): ReLU(inplace=True)
      (se): Identity()
      (conv3): Conv2d(72, 24, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn3): BatchNorm2d(24, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (act3): ReLU(inplace=True)
      (skip): Sequential(
        (0): Conv2d(16, 16, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), groups=16, bias=False)
        (1): BatchNorm2d(16, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
        (2): Conv2d(16, 24, kernel_size=(1, 1), stride=(1, 1))
        (3): BatchNorm2d(24, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      )
    )
    (2): Block(
      (conv1): Conv2d(24, 88, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn1): BatchNorm2d(88, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (act1): ReLU(inplace=True)
      (conv2): Conv2d(88, 88, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=88, bias=False)
      (bn2): BatchNorm2d(88, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (act2): ReLU(inplace=True)
      (se): Identity()
      (conv3): Conv2d(88, 24, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn3): BatchNorm2d(24, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (act3): ReLU(inplace=True)
    )
    (3): Block(
      (conv1): Conv2d(24, 96, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn1): BatchNorm2d(96, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (act1): Hardswish()
      (conv2): Conv2d(96, 96, kernel_size=(5, 5), stride=(2, 2), padding=(2, 2), groups=96, bias=False)
      (bn2): BatchNorm2d(96, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (act2): Hardswish()
      (se): SeModule(
        (se): Sequential(
          (0): AdaptiveAvgPool2d(output_size=1)
          (1): Conv2d(96, 24, kernel_size=(1, 1), stride=(1, 1), bias=False)
          (2): BatchNorm2d(24, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
          (3): ReLU(inplace=True)
          (4): Conv2d(24, 96, kernel_size=(1, 1), stride=(1, 1), bias=False)
          (5): Hardsigmoid()
        )
      )
      (conv3): Conv2d(96, 40, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn3): BatchNorm2d(40, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (act3): Hardswish()
      (skip): Sequential(
        (0): Conv2d(24, 24, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), groups=24, bias=False)
        (1): BatchNorm2d(24, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
        (2): Conv2d(24, 40, kernel_size=(1, 1), stride=(1, 1))
        (3): BatchNorm2d(40, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      )
    )
    (4): Block(
      (conv1): Conv2d(40, 240, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn1): BatchNorm2d(240, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (act1): Hardswish()
      (conv2): Conv2d(240, 240, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2), groups=240, bias=False)
      (bn2): BatchNorm2d(240, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (act2): Hardswish()
      (se): SeModule(
        (se): Sequential(
          (0): AdaptiveAvgPool2d(output_size=1)
          (1): Conv2d(240, 60, kernel_size=(1, 1), stride=(1, 1), bias=False)
          (2): BatchNorm2d(60, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
          (3): ReLU(inplace=True)
          (4): Conv2d(60, 240, kernel_size=(1, 1), stride=(1, 1), bias=False)
          (5): Hardsigmoid()
        )
      )
      (conv3): Conv2d(240, 40, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn3): BatchNorm2d(40, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (act3): Hardswish()
    )
    (5): Block(
      (conv1): Conv2d(40, 240, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn1): BatchNorm2d(240, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (act1): Hardswish()
      (conv2): Conv2d(240, 240, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2), groups=240, bias=False)
      (bn2): BatchNorm2d(240, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (act2): Hardswish()
      (se): SeModule(
        (se): Sequential(
          (0): AdaptiveAvgPool2d(output_size=1)
          (1): Conv2d(240, 60, kernel_size=(1, 1), stride=(1, 1), bias=False)
          (2): BatchNorm2d(60, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
          (3): ReLU(inplace=True)
          (4): Conv2d(60, 240, kernel_size=(1, 1), stride=(1, 1), bias=False)
          (5): Hardsigmoid()
        )
      )
      (conv3): Conv2d(240, 40, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn3): BatchNorm2d(40, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (act3): Hardswish()
    )
    (6): Block(
      (conv1): Conv2d(40, 120, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn1): BatchNorm2d(120, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (act1): Hardswish()
      (conv2): Conv2d(120, 120, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2), groups=120, bias=False)
      (bn2): BatchNorm2d(120, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (act2): Hardswish()
      (se): SeModule(
        (se): Sequential(
          (0): AdaptiveAvgPool2d(output_size=1)
          (1): Conv2d(120, 30, kernel_size=(1, 1), stride=(1, 1), bias=False)
          (2): BatchNorm2d(30, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
          (3): ReLU(inplace=True)
          (4): Conv2d(30, 120, kernel_size=(1, 1), stride=(1, 1), bias=False)
          (5): Hardsigmoid()
        )
      )
      (conv3): Conv2d(120, 48, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn3): BatchNorm2d(48, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (act3): Hardswish()
      (skip): Sequential(
        (0): Conv2d(40, 48, kernel_size=(1, 1), stride=(1, 1), bias=False)
        (1): BatchNorm2d(48, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      )
    )
    (7): Block(
      (conv1): Conv2d(48, 144, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn1): BatchNorm2d(144, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (act1): Hardswish()
      (conv2): Conv2d(144, 144, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2), groups=144, bias=False)
      (bn2): BatchNorm2d(144, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (act2): Hardswish()
      (se): SeModule(
        (se): Sequential(
          (0): AdaptiveAvgPool2d(output_size=1)
          (1): Conv2d(144, 36, kernel_size=(1, 1), stride=(1, 1), bias=False)
          (2): BatchNorm2d(36, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
          (3): ReLU(inplace=True)
          (4): Conv2d(36, 144, kernel_size=(1, 1), stride=(1, 1), bias=False)
          (5): Hardsigmoid()
        )
      )
      (conv3): Conv2d(144, 48, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn3): BatchNorm2d(48, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (act3): Hardswish()
    )
    (8): Block(
      (conv1): Conv2d(48, 288, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn1): BatchNorm2d(288, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (act1): Hardswish()
      (conv2): Conv2d(288, 288, kernel_size=(5, 5), stride=(2, 2), padding=(2, 2), groups=288, bias=False)
      (bn2): BatchNorm2d(288, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (act2): Hardswish()
      (se): SeModule(
        (se): Sequential(
          (0): AdaptiveAvgPool2d(output_size=1)
          (1): Conv2d(288, 72, kernel_size=(1, 1), stride=(1, 1), bias=False)
          (2): BatchNorm2d(72, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
          (3): ReLU(inplace=True)
          (4): Conv2d(72, 288, kernel_size=(1, 1), stride=(1, 1), bias=False)
          (5): Hardsigmoid()
        )
      )
      (conv3): Conv2d(288, 96, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn3): BatchNorm2d(96, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (act3): Hardswish()
      (skip): Sequential(
        (0): Conv2d(48, 48, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), groups=48, bias=False)
        (1): BatchNorm2d(48, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
        (2): Conv2d(48, 96, kernel_size=(1, 1), stride=(1, 1))
        (3): BatchNorm2d(96, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      )
    )
    (9): Block(
      (conv1): Conv2d(96, 576, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn1): BatchNorm2d(576, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (act1): Hardswish()
      (conv2): Conv2d(576, 576, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2), groups=576, bias=False)
      (bn2): BatchNorm2d(576, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (act2): Hardswish()
      (se): SeModule(
        (se): Sequential(
          (0): AdaptiveAvgPool2d(output_size=1)
          (1): Conv2d(576, 144, kernel_size=(1, 1), stride=(1, 1), bias=False)
          (2): BatchNorm2d(144, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
          (3): ReLU(inplace=True)
          (4): Conv2d(144, 576, kernel_size=(1, 1), stride=(1, 1), bias=False)
          (5): Hardsigmoid()
        )
      )
      (conv3): Conv2d(576, 96, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn3): BatchNorm2d(96, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (act3): Hardswish()
    )
    (10): Block(
      (conv1): Conv2d(96, 576, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn1): BatchNorm2d(576, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (act1): Hardswish()
      (conv2): Conv2d(576, 576, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2), groups=576, bias=False)
      (bn2): BatchNorm2d(576, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (act2): Hardswish()
      (se): SeModule(
        (se): Sequential(
          (0): AdaptiveAvgPool2d(output_size=1)
          (1): Conv2d(576, 144, kernel_size=(1, 1), stride=(1, 1), bias=False)
          (2): BatchNorm2d(144, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
          (3): ReLU(inplace=True)
          (4): Conv2d(144, 576, kernel_size=(1, 1), stride=(1, 1), bias=False)
          (5): Hardsigmoid()
        )
      )
      (conv3): Conv2d(576, 96, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn3): BatchNorm2d(96, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (act3): Hardswish()
    )
  )
  (conv2): Conv2d(96, 576, kernel_size=(1, 1), stride=(1, 1), bias=False)
  (bn2): BatchNorm2d(576, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
  (hs2): Hardswish()
  (gap): AdaptiveAvgPool2d(output_size=1)
  (linear3): Linear(in_features=576, out_features=1280, bias=False)
  (bn3): BatchNorm1d(1280, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
  (hs3): Hardswish()
  (drop): Dropout(p=0.2, inplace=False)
  (linear4): Linear(in_features=1280, out_features=1000, bias=True)
)
number of params: 2950524
LR = 0.00400000
Batch size = 4096
Update frequent = 2
Number of training examples = 1281167
Number of training training per epoch = 312
Param groups = {
  "decay": {
    "weight_decay": 0.05,
    "params": [
      "conv1.weight",
      "bneck.0.conv1.weight",
      "bneck.0.conv2.weight",
      "bneck.0.se.se.1.weight",
      "bneck.0.se.se.4.weight",
      "bneck.0.conv3.weight",
      "bneck.0.skip.0.weight",
      "bneck.1.conv1.weight",
      "bneck.1.conv2.weight",
      "bneck.1.conv3.weight",
      "bneck.1.skip.0.weight",
      "bneck.1.skip.2.weight",
      "bneck.2.conv1.weight",
      "bneck.2.conv2.weight",
      "bneck.2.conv3.weight",
      "bneck.3.conv1.weight",
      "bneck.3.conv2.weight",
      "bneck.3.se.se.1.weight",
      "bneck.3.se.se.4.weight",
      "bneck.3.conv3.weight",
      "bneck.3.skip.0.weight",
      "bneck.3.skip.2.weight",
      "bneck.4.conv1.weight",
      "bneck.4.conv2.weight",
      "bneck.4.se.se.1.weight",
      "bneck.4.se.se.4.weight",
      "bneck.4.conv3.weight",
      "bneck.5.conv1.weight",
      "bneck.5.conv2.weight",
      "bneck.5.se.se.1.weight",
      "bneck.5.se.se.4.weight",
      "bneck.5.conv3.weight",
      "bneck.6.conv1.weight",
      "bneck.6.conv2.weight",
      "bneck.6.se.se.1.weight",
      "bneck.6.se.se.4.weight",
      "bneck.6.conv3.weight",
      "bneck.6.skip.0.weight",
      "bneck.7.conv1.weight",
      "bneck.7.conv2.weight",
      "bneck.7.se.se.1.weight",
      "bneck.7.se.se.4.weight",
      "bneck.7.conv3.weight",
      "bneck.8.conv1.weight",
      "bneck.8.conv2.weight",
      "bneck.8.se.se.1.weight",
      "bneck.8.se.se.4.weight",
      "bneck.8.conv3.weight",
      "bneck.8.skip.0.weight",
      "bneck.8.skip.2.weight",
      "bneck.9.conv1.weight",
      "bneck.9.conv2.weight",
      "bneck.9.se.se.1.weight",
      "bneck.9.se.se.4.weight",
      "bneck.9.conv3.weight",
      "bneck.10.conv1.weight",
      "bneck.10.conv2.weight",
      "bneck.10.se.se.1.weight",
      "bneck.10.se.se.4.weight",
      "bneck.10.conv3.weight",
      "conv2.weight",
      "linear3.weight",
      "linear4.weight"
    ],
    "lr_scale": 1.0
  },
  "no_decay": {
    "weight_decay": 0.0,
    "params": [
      "bn1.weight",
      "bn1.bias",
      "bneck.0.bn1.weight",
      "bneck.0.bn1.bias",
      "bneck.0.bn2.weight",
      "bneck.0.bn2.bias",
      "bneck.0.se.se.2.weight",
      "bneck.0.se.se.2.bias",
      "bneck.0.bn3.weight",
      "bneck.0.bn3.bias",
      "bneck.0.skip.1.weight",
      "bneck.0.skip.1.bias",
      "bneck.1.bn1.weight",
      "bneck.1.bn1.bias",
      "bneck.1.bn2.weight",
      "bneck.1.bn2.bias",
      "bneck.1.bn3.weight",
      "bneck.1.bn3.bias",
      "bneck.1.skip.1.weight",
      "bneck.1.skip.1.bias",
      "bneck.1.skip.2.bias",
      "bneck.1.skip.3.weight",
      "bneck.1.skip.3.bias",
      "bneck.2.bn1.weight",
      "bneck.2.bn1.bias",
      "bneck.2.bn2.weight",
      "bneck.2.bn2.bias",
      "bneck.2.bn3.weight",
      "bneck.2.bn3.bias",
      "bneck.3.bn1.weight",
      "bneck.3.bn1.bias",
      "bneck.3.bn2.weight",
      "bneck.3.bn2.bias",
      "bneck.3.se.se.2.weight",
      "bneck.3.se.se.2.bias",
      "bneck.3.bn3.weight",
      "bneck.3.bn3.bias",
      "bneck.3.skip.1.weight",
      "bneck.3.skip.1.bias",
      "bneck.3.skip.2.bias",
      "bneck.3.skip.3.weight",
      "bneck.3.skip.3.bias",
      "bneck.4.bn1.weight",
      "bneck.4.bn1.bias",
      "bneck.4.bn2.weight",
      "bneck.4.bn2.bias",
      "bneck.4.se.se.2.weight",
      "bneck.4.se.se.2.bias",
      "bneck.4.bn3.weight",
      "bneck.4.bn3.bias",
      "bneck.5.bn1.weight",
      "bneck.5.bn1.bias",
      "bneck.5.bn2.weight",
      "bneck.5.bn2.bias",
      "bneck.5.se.se.2.weight",
      "bneck.5.se.se.2.bias",
      "bneck.5.bn3.weight",
      "bneck.5.bn3.bias",
      "bneck.6.bn1.weight",
      "bneck.6.bn1.bias",
      "bneck.6.bn2.weight",
      "bneck.6.bn2.bias",
      "bneck.6.se.se.2.weight",
      "bneck.6.se.se.2.bias",
      "bneck.6.bn3.weight",
      "bneck.6.bn3.bias",
      "bneck.6.skip.1.weight",
      "bneck.6.skip.1.bias",
      "bneck.7.bn1.weight",
      "bneck.7.bn1.bias",
      "bneck.7.bn2.weight",
      "bneck.7.bn2.bias",
      "bneck.7.se.se.2.weight",
      "bneck.7.se.se.2.bias",
      "bneck.7.bn3.weight",
      "bneck.7.bn3.bias",
      "bneck.8.bn1.weight",
      "bneck.8.bn1.bias",
      "bneck.8.bn2.weight",
      "bneck.8.bn2.bias",
      "bneck.8.se.se.2.weight",
      "bneck.8.se.se.2.bias",
      "bneck.8.bn3.weight",
      "bneck.8.bn3.bias",
      "bneck.8.skip.1.weight",
      "bneck.8.skip.1.bias",
      "bneck.8.skip.2.bias",
      "bneck.8.skip.3.weight",
      "bneck.8.skip.3.bias",
      "bneck.9.bn1.weight",
      "bneck.9.bn1.bias",
      "bneck.9.bn2.weight",
      "bneck.9.bn2.bias",
      "bneck.9.se.se.2.weight",
      "bneck.9.se.se.2.bias",
      "bneck.9.bn3.weight",
      "bneck.9.bn3.bias",
      "bneck.10.bn1.weight",
      "bneck.10.bn1.bias",
      "bneck.10.bn2.weight",
      "bneck.10.bn2.bias",
      "bneck.10.se.se.2.weight",
      "bneck.10.se.se.2.bias",
      "bneck.10.bn3.weight",
      "bneck.10.bn3.bias",
      "bn2.weight",
      "bn2.bias",
      "bn3.weight",
      "bn3.bias",
      "linear4.bias"
    ],
    "lr_scale": 1.0
  }
}
Use Cosine LR scheduler
Set warmup steps = 6240
Set warmup steps = 0
Max WD = 0.0500000, Min WD = 0.0500000
criterion = LabelSmoothingCrossEntropy()
Auto resume checkpoint: checkpoint/checkpoint-264.pth
Resume checkpoint checkpoint/checkpoint-264.pth
With optim & sched!
Start training for 300 epochs
Epoch: [265]  [  0/625]  eta: 5:12:11  lr: 0.000153  min_lr: 0.000153  loss: 2.6982 (2.6982)  class_acc: 0.6250 (0.6250)  weight_decay: 0.0500 (0.0500)  time: 29.9703  data: 22.2179  max mem: 2928
Epoch: [265]  [200/625]  eta: 0:15:33  lr: 0.000150  min_lr: 0.000150  loss: 2.7303 (2.7090)  class_acc: 0.5938 (0.5956)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.8810 (3.1684)  time: 1.8843  data: 0.0438  max mem: 2928
Epoch: [265]  [400/625]  eta: 0:08:04  lr: 0.000148  min_lr: 0.000148  loss: 2.7196 (2.7135)  class_acc: 0.5859 (0.5945)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.0067 (3.2041)  time: 2.1437  data: 0.3882  max mem: 2928
Epoch: [265]  [600/625]  eta: 0:00:52  lr: 0.000145  min_lr: 0.000145  loss: 2.7335 (2.7158)  class_acc: 0.5898 (0.5935)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7748 (3.3163)  time: 1.9821  data: 0.0006  max mem: 2928
Epoch: [265]  [624/625]  eta: 0:00:02  lr: 0.000145  min_lr: 0.000145  loss: 2.7468 (2.7158)  class_acc: 0.5977 (0.5937)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4024 (3.2918)  time: 0.7514  data: 0.0016  max mem: 2928
Epoch: [265] Total time: 0:21:43 (2.0851 s / it)
Averaged stats: lr: 0.000145  min_lr: 0.000145  loss: 2.7468 (2.7100)  class_acc: 0.5977 (0.5937)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4024 (3.2918)
Test:  [ 0/50]  eta: 0:10:16  loss: 1.4208 (1.4208)  acc1: 64.8000 (64.8000)  acc5: 89.6000 (89.6000)  time: 12.3266  data: 11.6150  max mem: 2928
Test:  [10/50]  eta: 0:01:23  loss: 1.3219 (1.3229)  acc1: 70.4000 (71.0545)  acc5: 89.6000 (89.3091)  time: 2.0756  data: 1.9935  max mem: 2928
Test:  [20/50]  eta: 0:00:50  loss: 1.4699 (1.4817)  acc1: 66.4000 (67.2381)  acc5: 88.0000 (87.8476)  time: 1.1575  data: 1.1392  max mem: 2928
Test:  [30/50]  eta: 0:00:30  loss: 1.6121 (1.4965)  acc1: 64.0000 (66.7097)  acc5: 87.2000 (87.3806)  time: 1.2436  data: 1.2263  max mem: 2928
Test:  [40/50]  eta: 0:00:13  loss: 1.5226 (1.5095)  acc1: 64.0000 (66.2634)  acc5: 86.4000 (87.1220)  time: 1.0960  data: 1.0781  max mem: 2928
Test:  [49/50]  eta: 0:00:01  loss: 1.4883 (1.5159)  acc1: 64.8000 (66.1120)  acc5: 86.4000 (86.9440)  time: 0.9563  data: 0.9368  max mem: 2928
Test: Total time: 0:00:58 (1.1724 s / it)
* Acc@1 67.028 Acc@5 86.856 loss 1.483
Accuracy of the model on the 50000 test images: 67.0%
Max accuracy: 67.03%
Epoch: [266]  [  0/625]  eta: 3:14:14  lr: 0.000145  min_lr: 0.000145  loss: 2.6366 (2.6366)  class_acc: 0.6133 (0.6133)  weight_decay: 0.0500 (0.0500)  time: 18.6478  data: 17.9495  max mem: 2928
Epoch: [266]  [200/625]  eta: 0:13:43  lr: 0.000142  min_lr: 0.000142  loss: 2.7148 (2.7098)  class_acc: 0.5898 (0.5929)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.2582 (3.2946)  time: 1.7768  data: 0.0008  max mem: 2928
Epoch: [266]  [400/625]  eta: 0:07:09  lr: 0.000139  min_lr: 0.000139  loss: 2.6743 (2.7173)  class_acc: 0.5938 (0.5913)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.8287 (3.3699)  time: 1.7315  data: 0.0008  max mem: 2928
Epoch: [266]  [600/625]  eta: 0:00:47  lr: 0.000137  min_lr: 0.000137  loss: 2.6792 (2.7112)  class_acc: 0.5938 (0.5925)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1729 (3.4514)  time: 1.9200  data: 0.0011  max mem: 2928
Epoch: [266]  [624/625]  eta: 0:00:01  lr: 0.000137  min_lr: 0.000137  loss: 2.7016 (2.7115)  class_acc: 0.6055 (0.5925)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.8838 (3.4796)  time: 0.8098  data: 0.0014  max mem: 2928
Epoch: [266] Total time: 0:19:27 (1.8684 s / it)
Averaged stats: lr: 0.000137  min_lr: 0.000137  loss: 2.7016 (2.7090)  class_acc: 0.6055 (0.5940)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.8838 (3.4796)
Test:  [ 0/50]  eta: 0:09:26  loss: 1.4158 (1.4158)  acc1: 68.0000 (68.0000)  acc5: 92.0000 (92.0000)  time: 11.3276  data: 11.2999  max mem: 2928
Test:  [10/50]  eta: 0:01:20  loss: 1.2840 (1.2876)  acc1: 72.0000 (72.6545)  acc5: 90.4000 (89.6000)  time: 2.0166  data: 1.9942  max mem: 2928
Test:  [20/50]  eta: 0:00:48  loss: 1.4544 (1.4495)  acc1: 68.0000 (68.4190)  acc5: 88.0000 (88.1524)  time: 1.1289  data: 1.1081  max mem: 2928
Test:  [30/50]  eta: 0:00:27  loss: 1.6032 (1.4796)  acc1: 64.0000 (67.0194)  acc5: 87.2000 (87.2516)  time: 1.0116  data: 0.9930  max mem: 2928
Test:  [40/50]  eta: 0:00:11  loss: 1.5103 (1.4976)  acc1: 63.2000 (66.4781)  acc5: 85.6000 (86.6146)  time: 0.5873  data: 0.5700  max mem: 2928
Test:  [49/50]  eta: 0:00:00  loss: 1.4662 (1.5046)  acc1: 63.2000 (65.9200)  acc5: 86.4000 (86.5280)  time: 0.4541  data: 0.4365  max mem: 2928
Test: Total time: 0:00:46 (0.9389 s / it)
* Acc@1 67.022 Acc@5 87.080 loss 1.470
Accuracy of the model on the 50000 test images: 67.0%
Max accuracy: 67.03%
Epoch: [267]  [  0/625]  eta: 3:20:25  lr: 0.000136  min_lr: 0.000136  loss: 2.8081 (2.8081)  class_acc: 0.5508 (0.5508)  weight_decay: 0.0500 (0.0500)  time: 19.2412  data: 16.5344  max mem: 2928
Epoch: [267]  [200/625]  eta: 0:13:44  lr: 0.000134  min_lr: 0.000134  loss: 2.7135 (2.7086)  class_acc: 0.6016 (0.5941)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2558 (inf)  time: 1.9102  data: 0.0138  max mem: 2928
Epoch: [267]  [400/625]  eta: 0:07:04  lr: 0.000131  min_lr: 0.000131  loss: 2.6886 (2.7125)  class_acc: 0.5898 (0.5919)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.9867 (inf)  time: 1.5592  data: 0.0006  max mem: 2928
Epoch: [267]  [600/625]  eta: 0:00:47  lr: 0.000129  min_lr: 0.000129  loss: 2.7519 (2.7126)  class_acc: 0.5898 (0.5923)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4030 (inf)  time: 1.8698  data: 0.0006  max mem: 2928
Epoch: [267]  [624/625]  eta: 0:00:01  lr: 0.000129  min_lr: 0.000129  loss: 2.7224 (2.7128)  class_acc: 0.5859 (0.5922)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.1920 (inf)  time: 0.7403  data: 0.0016  max mem: 2928
Epoch: [267] Total time: 0:19:13 (1.8462 s / it)
Averaged stats: lr: 0.000129  min_lr: 0.000129  loss: 2.7224 (2.7063)  class_acc: 0.5859 (0.5946)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.1920 (inf)
Test:  [ 0/50]  eta: 0:09:05  loss: 1.4205 (1.4205)  acc1: 66.4000 (66.4000)  acc5: 91.2000 (91.2000)  time: 10.9186  data: 10.8928  max mem: 2928
Test:  [10/50]  eta: 0:01:11  loss: 1.2860 (1.2785)  acc1: 71.2000 (71.7091)  acc5: 89.6000 (89.4545)  time: 1.7946  data: 1.7749  max mem: 2928
Test:  [20/50]  eta: 0:00:42  loss: 1.3355 (1.4082)  acc1: 68.0000 (68.6095)  acc5: 88.8000 (88.1143)  time: 0.9544  data: 0.9355  max mem: 2928
Test:  [30/50]  eta: 0:00:25  loss: 1.5094 (1.4340)  acc1: 65.6000 (67.7419)  acc5: 87.2000 (87.9226)  time: 0.9651  data: 0.9461  max mem: 2928
Test:  [40/50]  eta: 0:00:10  loss: 1.4308 (1.4614)  acc1: 65.6000 (67.2585)  acc5: 86.4000 (87.3171)  time: 0.7517  data: 0.7315  max mem: 2928
Test:  [49/50]  eta: 0:00:00  loss: 1.4840 (1.4681)  acc1: 66.4000 (67.0240)  acc5: 85.6000 (87.0720)  time: 0.5676  data: 0.5475  max mem: 2928
Test: Total time: 0:00:48 (0.9617 s / it)
* Acc@1 67.826 Acc@5 87.606 loss 1.430
Accuracy of the model on the 50000 test images: 67.8%
Max accuracy: 67.83%
Epoch: [268]  [  0/625]  eta: 3:28:20  lr: 0.000128  min_lr: 0.000128  loss: 2.5826 (2.5826)  class_acc: 0.6172 (0.6172)  weight_decay: 0.0500 (0.0500)  time: 20.0007  data: 17.7944  max mem: 2928
Epoch: [268]  [200/625]  eta: 0:13:31  lr: 0.000126  min_lr: 0.000126  loss: 2.7327 (2.7133)  class_acc: 0.5820 (0.5907)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.5326 (3.3510)  time: 1.7522  data: 0.0090  max mem: 2928
Epoch: [268]  [400/625]  eta: 0:07:00  lr: 0.000123  min_lr: 0.000123  loss: 2.6934 (2.7126)  class_acc: 0.6016 (0.5922)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.3752 (3.5178)  time: 1.9288  data: 0.0007  max mem: 2928
Epoch: [268]  [600/625]  eta: 0:00:47  lr: 0.000121  min_lr: 0.000121  loss: 2.7205 (2.7108)  class_acc: 0.5898 (0.5931)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.2686 (3.5203)  time: 2.0563  data: 0.0007  max mem: 2928
Epoch: [268]  [624/625]  eta: 0:00:01  lr: 0.000121  min_lr: 0.000121  loss: 2.6252 (2.7082)  class_acc: 0.6055 (0.5937)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.1555 (3.5092)  time: 0.9236  data: 0.0013  max mem: 2928
Epoch: [268] Total time: 0:19:12 (1.8445 s / it)
Averaged stats: lr: 0.000121  min_lr: 0.000121  loss: 2.6252 (2.7043)  class_acc: 0.6055 (0.5952)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.1555 (3.5092)
Test:  [ 0/50]  eta: 0:09:11  loss: 1.3064 (1.3064)  acc1: 64.8000 (64.8000)  acc5: 92.0000 (92.0000)  time: 11.0248  data: 10.9935  max mem: 2928
Test:  [10/50]  eta: 0:01:24  loss: 1.2414 (1.2534)  acc1: 72.8000 (72.3636)  acc5: 89.6000 (89.0909)  time: 2.1043  data: 2.0834  max mem: 2928
Test:  [20/50]  eta: 0:00:52  loss: 1.3990 (1.3931)  acc1: 68.0000 (68.3810)  acc5: 88.0000 (88.1524)  time: 1.2923  data: 1.2730  max mem: 2928
Test:  [30/50]  eta: 0:00:30  loss: 1.4889 (1.4157)  acc1: 65.6000 (67.8194)  acc5: 87.2000 (87.7677)  time: 1.2075  data: 1.1883  max mem: 2928
Test:  [40/50]  eta: 0:00:12  loss: 1.3815 (1.4375)  acc1: 65.6000 (67.4341)  acc5: 87.2000 (87.2976)  time: 0.7894  data: 0.7697  max mem: 2928
Test:  [49/50]  eta: 0:00:01  loss: 1.4162 (1.4396)  acc1: 65.6000 (67.1840)  acc5: 85.6000 (87.0240)  time: 0.5964  data: 0.5765  max mem: 2928
Test: Total time: 0:00:53 (1.0714 s / it)
* Acc@1 67.870 Acc@5 87.568 loss 1.408
Accuracy of the model on the 50000 test images: 67.9%
Max accuracy: 67.87%
Epoch: [269]  [  0/625]  eta: 3:56:16  lr: 0.000121  min_lr: 0.000121  loss: 2.5830 (2.5830)  class_acc: 0.6328 (0.6328)  weight_decay: 0.0500 (0.0500)  time: 22.6824  data: 15.8010  max mem: 2928
Epoch: [269]  [200/625]  eta: 0:14:00  lr: 0.000118  min_lr: 0.000118  loss: 2.7129 (2.7029)  class_acc: 0.6133 (0.5974)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7937 (3.0471)  time: 1.9056  data: 0.0009  max mem: 2928
Epoch: [269]  [400/625]  eta: 0:07:15  lr: 0.000116  min_lr: 0.000116  loss: 2.7001 (2.7024)  class_acc: 0.5898 (0.5967)  weight_decay: 0.0500 (0.0500)  grad_norm: 4.0177 (3.2485)  time: 1.9665  data: 0.0010  max mem: 2928
Epoch: [269]  [600/625]  eta: 0:00:48  lr: 0.000113  min_lr: 0.000113  loss: 2.7637 (2.7064)  class_acc: 0.5820 (0.5955)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.1701 (3.4919)  time: 1.9778  data: 0.0008  max mem: 2928
Epoch: [269]  [624/625]  eta: 0:00:01  lr: 0.000113  min_lr: 0.000113  loss: 2.7322 (2.7067)  class_acc: 0.5898 (0.5953)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.7104 (3.5100)  time: 0.7780  data: 0.0014  max mem: 2928
Epoch: [269] Total time: 0:19:35 (1.8811 s / it)
Averaged stats: lr: 0.000113  min_lr: 0.000113  loss: 2.7322 (2.7040)  class_acc: 0.5898 (0.5953)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.7104 (3.5100)
Test:  [ 0/50]  eta: 0:10:13  loss: 1.4209 (1.4209)  acc1: 67.2000 (67.2000)  acc5: 88.0000 (88.0000)  time: 12.2786  data: 12.2494  max mem: 2928
Test:  [10/50]  eta: 0:01:16  loss: 1.2628 (1.2641)  acc1: 71.2000 (72.1455)  acc5: 90.4000 (89.3818)  time: 1.9220  data: 1.9020  max mem: 2928
Test:  [20/50]  eta: 0:00:45  loss: 1.4259 (1.4152)  acc1: 68.8000 (69.0667)  acc5: 88.0000 (88.0381)  time: 0.9817  data: 0.9617  max mem: 2928
Test:  [30/50]  eta: 0:00:27  loss: 1.5233 (1.4392)  acc1: 64.0000 (68.0258)  acc5: 87.2000 (87.7419)  time: 1.0510  data: 1.0297  max mem: 2928
Test:  [40/50]  eta: 0:00:11  loss: 1.4580 (1.4621)  acc1: 64.0000 (67.5122)  acc5: 87.2000 (87.2781)  time: 0.7457  data: 0.7250  max mem: 2928
Test:  [49/50]  eta: 0:00:00  loss: 1.4548 (1.4639)  acc1: 64.0000 (67.2320)  acc5: 87.2000 (87.1200)  time: 0.6696  data: 0.6486  max mem: 2928
Test: Total time: 0:00:48 (0.9767 s / it)
* Acc@1 67.794 Acc@5 87.504 loss 1.426
Accuracy of the model on the 50000 test images: 67.8%
Max accuracy: 67.87%
Epoch: [270]  [  0/625]  eta: 3:36:47  lr: 0.000113  min_lr: 0.000113  loss: 2.6511 (2.6511)  class_acc: 0.6211 (0.6211)  weight_decay: 0.0500 (0.0500)  time: 20.8115  data: 19.1630  max mem: 2928
Epoch: [270]  [200/625]  eta: 0:13:58  lr: 0.000111  min_lr: 0.000111  loss: 2.6846 (2.7009)  class_acc: 0.5898 (0.5976)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5810 (2.8145)  time: 1.9970  data: 0.0596  max mem: 2928
Epoch: [270]  [400/625]  eta: 0:07:13  lr: 0.000109  min_lr: 0.000109  loss: 2.7432 (2.7050)  class_acc: 0.5938 (0.5966)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2566 (2.9775)  time: 1.7794  data: 0.0006  max mem: 2928
Epoch: [270]  [600/625]  eta: 0:00:48  lr: 0.000106  min_lr: 0.000106  loss: 2.6925 (2.7024)  class_acc: 0.5898 (0.5963)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7629 (3.0018)  time: 1.8598  data: 0.0007  max mem: 2928
Epoch: [270]  [624/625]  eta: 0:00:01  lr: 0.000106  min_lr: 0.000106  loss: 2.6738 (2.7020)  class_acc: 0.6016 (0.5965)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.2650 (3.0529)  time: 0.7898  data: 0.0013  max mem: 2928
Epoch: [270] Total time: 0:19:39 (1.8865 s / it)
Averaged stats: lr: 0.000106  min_lr: 0.000106  loss: 2.6738 (2.7029)  class_acc: 0.6016 (0.5959)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.2650 (3.0529)
Test:  [ 0/50]  eta: 0:09:59  loss: 1.3246 (1.3246)  acc1: 68.0000 (68.0000)  acc5: 88.8000 (88.8000)  time: 11.9856  data: 11.9541  max mem: 2928
Test:  [10/50]  eta: 0:01:22  loss: 1.2602 (1.2566)  acc1: 72.0000 (71.9273)  acc5: 90.4000 (89.9636)  time: 2.0617  data: 2.0425  max mem: 2928
Test:  [20/50]  eta: 0:00:50  loss: 1.4536 (1.4100)  acc1: 67.2000 (68.3429)  acc5: 88.0000 (88.3048)  time: 1.1678  data: 1.1491  max mem: 2928
Test:  [30/50]  eta: 0:00:28  loss: 1.5538 (1.4361)  acc1: 64.8000 (67.6903)  acc5: 86.4000 (87.6645)  time: 1.0898  data: 1.0703  max mem: 2928
Test:  [40/50]  eta: 0:00:11  loss: 1.4390 (1.4523)  acc1: 64.8000 (67.4146)  acc5: 86.4000 (87.3561)  time: 0.6133  data: 0.5933  max mem: 2928
Test:  [49/50]  eta: 0:00:00  loss: 1.4204 (1.4527)  acc1: 64.8000 (67.0720)  acc5: 87.2000 (87.2160)  time: 0.4896  data: 0.4698  max mem: 2928
Test: Total time: 0:00:48 (0.9752 s / it)
* Acc@1 67.814 Acc@5 87.684 loss 1.412
Accuracy of the model on the 50000 test images: 67.8%
Max accuracy: 67.87%
Epoch: [271]  [  0/625]  eta: 3:39:25  lr: 0.000106  min_lr: 0.000106  loss: 2.7391 (2.7391)  class_acc: 0.5820 (0.5820)  weight_decay: 0.0500 (0.0500)  time: 21.0640  data: 16.7909  max mem: 2928
Epoch: [271]  [200/625]  eta: 0:14:18  lr: 0.000104  min_lr: 0.000104  loss: 2.7004 (2.6886)  class_acc: 0.5898 (0.5973)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.9814 (3.3375)  time: 1.9515  data: 0.0007  max mem: 2928
Epoch: [271]  [400/625]  eta: 0:07:17  lr: 0.000101  min_lr: 0.000101  loss: 2.7393 (2.6974)  class_acc: 0.5820 (0.5966)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.8755 (3.3907)  time: 1.8603  data: 0.0008  max mem: 2928
Epoch: [271]  [600/625]  eta: 0:00:48  lr: 0.000099  min_lr: 0.000099  loss: 2.6825 (2.6988)  class_acc: 0.5859 (0.5960)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.3325 (3.2279)  time: 2.0133  data: 0.0007  max mem: 2928
Epoch: [271]  [624/625]  eta: 0:00:01  lr: 0.000099  min_lr: 0.000099  loss: 2.6910 (2.6987)  class_acc: 0.5898 (0.5962)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6265 (3.2198)  time: 0.6447  data: 0.0015  max mem: 2928
Epoch: [271] Total time: 0:19:36 (1.8828 s / it)
Averaged stats: lr: 0.000099  min_lr: 0.000099  loss: 2.6910 (2.6998)  class_acc: 0.5898 (0.5961)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6265 (3.2198)
Test:  [ 0/50]  eta: 0:09:45  loss: 1.4222 (1.4222)  acc1: 65.6000 (65.6000)  acc5: 89.6000 (89.6000)  time: 11.7017  data: 11.6789  max mem: 2928
Test:  [10/50]  eta: 0:01:18  loss: 1.2510 (1.2580)  acc1: 72.8000 (71.8545)  acc5: 90.4000 (90.4000)  time: 1.9592  data: 1.9403  max mem: 2928
Test:  [20/50]  eta: 0:00:46  loss: 1.3864 (1.4104)  acc1: 65.6000 (68.1524)  acc5: 88.8000 (88.8000)  time: 1.0353  data: 1.0157  max mem: 2928
Test:  [30/50]  eta: 0:00:27  loss: 1.4963 (1.4314)  acc1: 65.6000 (67.8710)  acc5: 88.0000 (88.3613)  time: 1.0415  data: 1.0215  max mem: 2928
Test:  [40/50]  eta: 0:00:11  loss: 1.4401 (1.4553)  acc1: 65.6000 (67.2000)  acc5: 86.4000 (87.7073)  time: 0.7603  data: 0.7414  max mem: 2928
Test:  [49/50]  eta: 0:00:00  loss: 1.4635 (1.4612)  acc1: 65.6000 (66.8960)  acc5: 85.6000 (87.4400)  time: 0.6726  data: 0.6535  max mem: 2928
Test: Total time: 0:00:48 (0.9737 s / it)
* Acc@1 67.910 Acc@5 87.666 loss 1.427
Accuracy of the model on the 50000 test images: 67.9%
Max accuracy: 67.91%
Epoch: [272]  [  0/625]  eta: 3:11:09  lr: 0.000099  min_lr: 0.000099  loss: 2.7360 (2.7360)  class_acc: 0.5977 (0.5977)  weight_decay: 0.0500 (0.0500)  time: 18.3508  data: 18.0127  max mem: 2928
Epoch: [272]  [200/625]  eta: 0:13:39  lr: 0.000097  min_lr: 0.000097  loss: 2.6547 (2.7008)  class_acc: 0.6016 (0.5970)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5486 (3.1072)  time: 1.8958  data: 1.1827  max mem: 2928
Epoch: [272]  [400/625]  eta: 0:07:00  lr: 0.000094  min_lr: 0.000094  loss: 2.6620 (2.6965)  class_acc: 0.5938 (0.5983)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3547 (3.1378)  time: 1.8289  data: 0.0006  max mem: 2928
Epoch: [272]  [600/625]  eta: 0:00:47  lr: 0.000092  min_lr: 0.000092  loss: 2.6817 (2.6969)  class_acc: 0.5859 (0.5973)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2615 (3.1892)  time: 1.9939  data: 0.0005  max mem: 2928
Epoch: [272]  [624/625]  eta: 0:00:01  lr: 0.000092  min_lr: 0.000092  loss: 2.6530 (2.6974)  class_acc: 0.5898 (0.5969)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5169 (3.1918)  time: 0.6706  data: 0.0014  max mem: 2928
Epoch: [272] Total time: 0:19:23 (1.8611 s / it)
Averaged stats: lr: 0.000092  min_lr: 0.000092  loss: 2.6530 (2.6984)  class_acc: 0.5898 (0.5969)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5169 (3.1918)
Test:  [ 0/50]  eta: 0:09:22  loss: 1.3835 (1.3835)  acc1: 64.0000 (64.0000)  acc5: 89.6000 (89.6000)  time: 11.2520  data: 11.2204  max mem: 2928
Test:  [10/50]  eta: 0:01:15  loss: 1.2347 (1.2630)  acc1: 72.8000 (72.2909)  acc5: 89.6000 (89.8909)  time: 1.8942  data: 1.8740  max mem: 2928
Test:  [20/50]  eta: 0:00:44  loss: 1.4162 (1.3927)  acc1: 68.8000 (68.9905)  acc5: 88.8000 (88.5333)  time: 1.0050  data: 0.9854  max mem: 2928
Test:  [30/50]  eta: 0:00:26  loss: 1.5001 (1.4099)  acc1: 66.4000 (68.4903)  acc5: 86.4000 (88.1548)  time: 1.0020  data: 0.9819  max mem: 2928
Test:  [40/50]  eta: 0:00:11  loss: 1.3811 (1.4346)  acc1: 66.4000 (67.7268)  acc5: 86.4000 (87.5317)  time: 0.7838  data: 0.7626  max mem: 2928
Test:  [49/50]  eta: 0:00:01  loss: 1.4691 (1.4392)  acc1: 65.6000 (67.3600)  acc5: 86.4000 (87.3440)  time: 0.7103  data: 0.6904  max mem: 2928
Test: Total time: 0:00:50 (1.0171 s / it)
* Acc@1 68.182 Acc@5 87.624 loss 1.409
Accuracy of the model on the 50000 test images: 68.2%
Max accuracy: 68.18%
Epoch: [273]  [  0/625]  eta: 3:22:31  lr: 0.000092  min_lr: 0.000092  loss: 2.5506 (2.5506)  class_acc: 0.6211 (0.6211)  weight_decay: 0.0500 (0.0500)  time: 19.4418  data: 19.2721  max mem: 2928
Epoch: [273]  [200/625]  eta: 0:14:02  lr: 0.000090  min_lr: 0.000090  loss: 2.6592 (2.6863)  class_acc: 0.6055 (0.6004)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7570 (3.0974)  time: 1.7972  data: 0.0008  max mem: 2928
Epoch: [273]  [400/625]  eta: 0:07:12  lr: 0.000088  min_lr: 0.000088  loss: 2.6879 (2.6899)  class_acc: 0.6016 (0.5990)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.9832 (3.2100)  time: 1.8369  data: 0.0007  max mem: 2928
Epoch: [273]  [600/625]  eta: 0:00:48  lr: 0.000086  min_lr: 0.000086  loss: 2.6857 (2.6927)  class_acc: 0.5938 (0.5985)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.8768 (inf)  time: 1.8753  data: 0.8175  max mem: 2928
Epoch: [273]  [624/625]  eta: 0:00:01  lr: 0.000085  min_lr: 0.000085  loss: 2.6485 (2.6922)  class_acc: 0.6094 (0.5987)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4341 (inf)  time: 0.9818  data: 0.5894  max mem: 2928
Epoch: [273] Total time: 0:19:38 (1.8858 s / it)
Averaged stats: lr: 0.000085  min_lr: 0.000085  loss: 2.6485 (2.6966)  class_acc: 0.6094 (0.5969)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4341 (inf)
Test:  [ 0/50]  eta: 0:10:09  loss: 1.3885 (1.3885)  acc1: 68.0000 (68.0000)  acc5: 89.6000 (89.6000)  time: 12.1982  data: 12.1731  max mem: 2928
Test:  [10/50]  eta: 0:01:21  loss: 1.2585 (1.2638)  acc1: 72.8000 (72.7273)  acc5: 90.4000 (89.9636)  time: 2.0263  data: 2.0053  max mem: 2928
Test:  [20/50]  eta: 0:00:48  loss: 1.3941 (1.4129)  acc1: 67.2000 (69.0667)  acc5: 88.0000 (88.3810)  time: 1.0749  data: 1.0555  max mem: 2928
Test:  [30/50]  eta: 0:00:27  loss: 1.5349 (1.4345)  acc1: 65.6000 (68.3097)  acc5: 88.0000 (87.8452)  time: 1.0214  data: 1.0026  max mem: 2928
Test:  [40/50]  eta: 0:00:12  loss: 1.4251 (1.4469)  acc1: 65.6000 (67.8829)  acc5: 87.2000 (87.4537)  time: 0.7823  data: 0.7636  max mem: 2928
Test:  [49/50]  eta: 0:00:01  loss: 1.4251 (1.4510)  acc1: 64.8000 (67.4560)  acc5: 87.2000 (87.3600)  time: 0.7817  data: 0.7629  max mem: 2928
Test: Total time: 0:00:51 (1.0345 s / it)
* Acc@1 68.180 Acc@5 87.878 loss 1.417
Accuracy of the model on the 50000 test images: 68.2%
Max accuracy: 68.18%
Epoch: [274]  [  0/625]  eta: 4:00:35  lr: 0.000085  min_lr: 0.000085  loss: 2.6856 (2.6856)  class_acc: 0.6289 (0.6289)  weight_decay: 0.0500 (0.0500)  time: 23.0969  data: 19.3993  max mem: 2928
Epoch: [274]  [200/625]  eta: 0:13:41  lr: 0.000083  min_lr: 0.000083  loss: 2.6918 (2.6992)  class_acc: 0.5859 (0.5967)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7592 (3.1120)  time: 1.9756  data: 0.0820  max mem: 2928
Epoch: [274]  [400/625]  eta: 0:07:03  lr: 0.000081  min_lr: 0.000081  loss: 2.6703 (2.6947)  class_acc: 0.6133 (0.5979)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6652 (3.0480)  time: 1.8408  data: 0.4873  max mem: 2928
Epoch: [274]  [600/625]  eta: 0:00:47  lr: 0.000079  min_lr: 0.000079  loss: 2.7090 (2.6946)  class_acc: 0.6016 (0.5976)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6876 (3.1544)  time: 1.9572  data: 0.0007  max mem: 2928
Epoch: [274]  [624/625]  eta: 0:00:01  lr: 0.000079  min_lr: 0.000079  loss: 2.6884 (2.6948)  class_acc: 0.5938 (0.5977)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2495 (3.1419)  time: 0.7670  data: 0.0015  max mem: 2928
Epoch: [274] Total time: 0:19:33 (1.8773 s / it)
Averaged stats: lr: 0.000079  min_lr: 0.000079  loss: 2.6884 (2.6941)  class_acc: 0.5938 (0.5972)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2495 (3.1419)
Test:  [ 0/50]  eta: 0:10:05  loss: 1.3563 (1.3563)  acc1: 67.2000 (67.2000)  acc5: 88.8000 (88.8000)  time: 12.1057  data: 12.0772  max mem: 2928
Test:  [10/50]  eta: 0:01:23  loss: 1.2730 (1.2644)  acc1: 70.4000 (71.9273)  acc5: 90.4000 (89.6000)  time: 2.0841  data: 2.0629  max mem: 2928
Test:  [20/50]  eta: 0:00:48  loss: 1.3975 (1.4028)  acc1: 68.8000 (68.7238)  acc5: 89.6000 (88.4952)  time: 1.1051  data: 1.0849  max mem: 2928
Test:  [30/50]  eta: 0:00:27  loss: 1.5271 (1.4267)  acc1: 66.4000 (67.8968)  acc5: 87.2000 (88.0000)  time: 1.0038  data: 0.9830  max mem: 2928
Test:  [40/50]  eta: 0:00:11  loss: 1.4284 (1.4394)  acc1: 66.4000 (67.6293)  acc5: 86.4000 (87.5902)  time: 0.6486  data: 0.6277  max mem: 2928
Test:  [49/50]  eta: 0:00:00  loss: 1.4388 (1.4441)  acc1: 65.6000 (67.2000)  acc5: 86.4000 (87.3920)  time: 0.6400  data: 0.6200  max mem: 2928
Test: Total time: 0:00:48 (0.9751 s / it)
* Acc@1 68.150 Acc@5 87.822 loss 1.410
Accuracy of the model on the 50000 test images: 68.2%
Max accuracy: 68.18%
Epoch: [275]  [  0/625]  eta: 3:53:07  lr: 0.000079  min_lr: 0.000079  loss: 2.8161 (2.8161)  class_acc: 0.5703 (0.5703)  weight_decay: 0.0500 (0.0500)  time: 22.3797  data: 13.6702  max mem: 2928
Epoch: [275]  [200/625]  eta: 0:14:00  lr: 0.000077  min_lr: 0.000077  loss: 2.7157 (2.6953)  class_acc: 0.5898 (0.5984)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.4296 (3.4390)  time: 1.8427  data: 0.0006  max mem: 2928
Epoch: [275]  [400/625]  eta: 0:07:12  lr: 0.000075  min_lr: 0.000075  loss: 2.7074 (2.6958)  class_acc: 0.5898 (0.5973)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5290 (3.3259)  time: 1.8275  data: 0.0006  max mem: 2928
Epoch: [275]  [600/625]  eta: 0:00:47  lr: 0.000073  min_lr: 0.000073  loss: 2.7033 (2.7010)  class_acc: 0.5820 (0.5961)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.9741 (3.2677)  time: 2.0943  data: 0.0006  max mem: 2928
Epoch: [275]  [624/625]  eta: 0:00:01  lr: 0.000073  min_lr: 0.000073  loss: 2.6872 (2.7010)  class_acc: 0.5859 (0.5957)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.2051 (3.2667)  time: 0.8218  data: 0.0013  max mem: 2928
Epoch: [275] Total time: 0:19:29 (1.8706 s / it)
Averaged stats: lr: 0.000073  min_lr: 0.000073  loss: 2.6872 (2.6949)  class_acc: 0.5859 (0.5975)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.2051 (3.2667)
Test:  [ 0/50]  eta: 0:10:11  loss: 1.3697 (1.3697)  acc1: 64.0000 (64.0000)  acc5: 90.4000 (90.4000)  time: 12.2342  data: 12.2045  max mem: 2928
Test:  [10/50]  eta: 0:01:20  loss: 1.2814 (1.2650)  acc1: 72.8000 (71.8545)  acc5: 90.4000 (89.3818)  time: 2.0091  data: 1.9897  max mem: 2928
Test:  [20/50]  eta: 0:00:46  loss: 1.3902 (1.4106)  acc1: 68.0000 (68.6857)  acc5: 88.0000 (88.1143)  time: 1.0152  data: 0.9965  max mem: 2928
Test:  [30/50]  eta: 0:00:27  loss: 1.5326 (1.4249)  acc1: 66.4000 (68.1806)  acc5: 87.2000 (87.7936)  time: 1.0058  data: 0.9864  max mem: 2928
Test:  [40/50]  eta: 0:00:11  loss: 1.3861 (1.4402)  acc1: 66.4000 (67.6683)  acc5: 87.2000 (87.5317)  time: 0.8079  data: 0.7866  max mem: 2928
Test:  [49/50]  eta: 0:00:01  loss: 1.4318 (1.4417)  acc1: 65.6000 (67.5040)  acc5: 87.2000 (87.3280)  time: 0.7010  data: 0.6800  max mem: 2928
Test: Total time: 0:00:51 (1.0329 s / it)
* Acc@1 68.200 Acc@5 87.736 loss 1.410
Accuracy of the model on the 50000 test images: 68.2%
Max accuracy: 68.20%
Epoch: [276]  [  0/625]  eta: 3:17:13  lr: 0.000073  min_lr: 0.000073  loss: 2.6518 (2.6518)  class_acc: 0.6250 (0.6250)  weight_decay: 0.0500 (0.0500)  time: 18.9342  data: 17.9040  max mem: 2928
Epoch: [276]  [200/625]  eta: 0:14:00  lr: 0.000071  min_lr: 0.000071  loss: 2.6926 (2.6846)  class_acc: 0.6016 (0.5994)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.7429 (3.3839)  time: 1.8723  data: 0.0054  max mem: 2928
Epoch: [276]  [400/625]  eta: 0:07:13  lr: 0.000069  min_lr: 0.000069  loss: 2.6608 (2.6845)  class_acc: 0.6094 (0.6008)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.1121 (3.4721)  time: 1.8420  data: 0.0010  max mem: 2928
Epoch: [276]  [600/625]  eta: 0:00:48  lr: 0.000067  min_lr: 0.000067  loss: 2.6647 (2.6884)  class_acc: 0.5977 (0.5993)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.9250 (3.2657)  time: 1.9726  data: 0.0007  max mem: 2928
Epoch: [276]  [624/625]  eta: 0:00:01  lr: 0.000067  min_lr: 0.000067  loss: 2.7245 (2.6897)  class_acc: 0.5938 (0.5994)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.9789 (3.2508)  time: 0.7098  data: 0.0016  max mem: 2928
Epoch: [276] Total time: 0:19:38 (1.8863 s / it)
Averaged stats: lr: 0.000067  min_lr: 0.000067  loss: 2.7245 (2.6929)  class_acc: 0.5938 (0.5978)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.9789 (3.2508)
Test:  [ 0/50]  eta: 0:10:29  loss: 1.3327 (1.3327)  acc1: 68.0000 (68.0000)  acc5: 92.0000 (92.0000)  time: 12.5927  data: 12.5655  max mem: 2928
Test:  [10/50]  eta: 0:01:23  loss: 1.2554 (1.2507)  acc1: 72.0000 (72.5091)  acc5: 90.4000 (90.1091)  time: 2.0772  data: 2.0560  max mem: 2928
Test:  [20/50]  eta: 0:00:48  loss: 1.4097 (1.4020)  acc1: 68.0000 (68.7619)  acc5: 89.6000 (88.7238)  time: 1.0580  data: 1.0379  max mem: 2928
Test:  [30/50]  eta: 0:00:28  loss: 1.5303 (1.4193)  acc1: 66.4000 (68.0774)  acc5: 88.0000 (88.2581)  time: 1.0705  data: 1.0514  max mem: 2928
Test:  [40/50]  eta: 0:00:12  loss: 1.4000 (1.4330)  acc1: 66.4000 (67.6683)  acc5: 87.2000 (87.8829)  time: 0.8149  data: 0.7966  max mem: 2928
Test:  [49/50]  eta: 0:00:01  loss: 1.4106 (1.4365)  acc1: 66.4000 (67.4400)  acc5: 87.2000 (87.6800)  time: 0.7385  data: 0.7200  max mem: 2928
Test: Total time: 0:00:50 (1.0160 s / it)
* Acc@1 68.428 Acc@5 87.984 loss 1.402
Accuracy of the model on the 50000 test images: 68.4%
Max accuracy: 68.43%
Epoch: [277]  [  0/625]  eta: 3:31:49  lr: 0.000067  min_lr: 0.000067  loss: 2.6082 (2.6082)  class_acc: 0.5859 (0.5859)  weight_decay: 0.0500 (0.0500)  time: 20.3358  data: 17.3456  max mem: 2928
Epoch: [277]  [200/625]  eta: 0:14:25  lr: 0.000065  min_lr: 0.000065  loss: 2.6370 (2.6730)  class_acc: 0.5938 (0.5995)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.3387 (3.1867)  time: 1.9612  data: 0.0012  max mem: 2928
Epoch: [277]  [400/625]  eta: 0:07:20  lr: 0.000064  min_lr: 0.000064  loss: 2.6703 (2.6893)  class_acc: 0.5977 (0.5964)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7297 (3.4342)  time: 1.8914  data: 0.0014  max mem: 2928
Epoch: [277]  [600/625]  eta: 0:00:48  lr: 0.000062  min_lr: 0.000062  loss: 2.7367 (2.6900)  class_acc: 0.5820 (0.5967)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.8013 (3.2965)  time: 1.9486  data: 0.0009  max mem: 2928
Epoch: [277]  [624/625]  eta: 0:00:01  lr: 0.000062  min_lr: 0.000062  loss: 2.7087 (2.6901)  class_acc: 0.5898 (0.5967)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.8374 (3.3087)  time: 0.7608  data: 0.0020  max mem: 2928
Epoch: [277] Total time: 0:19:45 (1.8971 s / it)
Averaged stats: lr: 0.000062  min_lr: 0.000062  loss: 2.7087 (2.6914)  class_acc: 0.5898 (0.5981)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.8374 (3.3087)
Test:  [ 0/50]  eta: 0:10:28  loss: 1.4060 (1.4060)  acc1: 65.6000 (65.6000)  acc5: 88.8000 (88.8000)  time: 12.5743  data: 12.5466  max mem: 2928
Test:  [10/50]  eta: 0:01:24  loss: 1.2714 (1.2757)  acc1: 73.6000 (72.6545)  acc5: 90.4000 (89.6727)  time: 2.1007  data: 2.0786  max mem: 2928
Test:  [20/50]  eta: 0:00:48  loss: 1.4380 (1.4259)  acc1: 68.0000 (68.7619)  acc5: 88.0000 (88.2286)  time: 1.0704  data: 1.0505  max mem: 2928
Test:  [30/50]  eta: 0:00:28  loss: 1.5226 (1.4390)  acc1: 65.6000 (68.0774)  acc5: 87.2000 (87.8194)  time: 1.0493  data: 1.0311  max mem: 2928
Test:  [40/50]  eta: 0:00:11  loss: 1.4300 (1.4517)  acc1: 66.4000 (67.4927)  acc5: 88.0000 (87.6488)  time: 0.7532  data: 0.7349  max mem: 2928
Test:  [49/50]  eta: 0:00:01  loss: 1.4326 (1.4545)  acc1: 64.8000 (67.2800)  acc5: 87.2000 (87.4720)  time: 0.7197  data: 0.7012  max mem: 2928
Test: Total time: 0:00:50 (1.0183 s / it)
* Acc@1 68.194 Acc@5 87.870 loss 1.414
Accuracy of the model on the 50000 test images: 68.2%
Max accuracy: 68.43%
Epoch: [278]  [  0/625]  eta: 4:17:04  lr: 0.000062  min_lr: 0.000062  loss: 2.7227 (2.7227)  class_acc: 0.6250 (0.6250)  weight_decay: 0.0500 (0.0500)  time: 24.6787  data: 16.5646  max mem: 2928
Epoch: [278]  [200/625]  eta: 0:14:11  lr: 0.000060  min_lr: 0.000060  loss: 2.6780 (2.7018)  class_acc: 0.5859 (0.5976)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.2220 (3.2738)  time: 1.8835  data: 0.0058  max mem: 2928
Epoch: [278]  [400/625]  eta: 0:07:18  lr: 0.000058  min_lr: 0.000058  loss: 2.6360 (2.6924)  class_acc: 0.6016 (0.6000)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7605 (3.2574)  time: 1.9273  data: 0.0011  max mem: 2928
Epoch: [278]  [600/625]  eta: 0:00:48  lr: 0.000056  min_lr: 0.000056  loss: 2.6787 (2.6932)  class_acc: 0.5938 (0.5991)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.0789 (3.1963)  time: 2.0496  data: 0.0011  max mem: 2928
Epoch: [278]  [624/625]  eta: 0:00:01  lr: 0.000056  min_lr: 0.000056  loss: 2.6510 (2.6931)  class_acc: 0.6055 (0.5991)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.8214 (3.2008)  time: 0.7820  data: 0.0018  max mem: 2928
Epoch: [278] Total time: 0:19:50 (1.9055 s / it)
Averaged stats: lr: 0.000056  min_lr: 0.000056  loss: 2.6510 (2.6903)  class_acc: 0.6055 (0.5988)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.8214 (3.2008)
Test:  [ 0/50]  eta: 0:10:03  loss: 1.3870 (1.3870)  acc1: 67.2000 (67.2000)  acc5: 90.4000 (90.4000)  time: 12.0789  data: 12.0518  max mem: 2928
Test:  [10/50]  eta: 0:01:22  loss: 1.2373 (1.2620)  acc1: 72.8000 (72.8727)  acc5: 90.4000 (89.9636)  time: 2.0694  data: 2.0496  max mem: 2928
Test:  [20/50]  eta: 0:00:49  loss: 1.4269 (1.4167)  acc1: 68.0000 (68.9524)  acc5: 88.0000 (88.5333)  time: 1.1394  data: 1.1205  max mem: 2928
Test:  [30/50]  eta: 0:00:29  loss: 1.5423 (1.4308)  acc1: 66.4000 (68.1806)  acc5: 87.2000 (88.2323)  time: 1.1569  data: 1.1388  max mem: 2928
Test:  [40/50]  eta: 0:00:12  loss: 1.4451 (1.4495)  acc1: 66.4000 (67.7073)  acc5: 87.2000 (87.8244)  time: 0.8331  data: 0.8144  max mem: 2928
Test:  [49/50]  eta: 0:00:01  loss: 1.4349 (1.4516)  acc1: 66.4000 (67.6160)  acc5: 87.2000 (87.6000)  time: 0.7899  data: 0.7698  max mem: 2928
Test: Total time: 0:00:52 (1.0488 s / it)
* Acc@1 68.216 Acc@5 87.900 loss 1.417
Accuracy of the model on the 50000 test images: 68.2%
Max accuracy: 68.43%
Epoch: [279]  [  0/625]  eta: 3:28:48  lr: 0.000056  min_lr: 0.000056  loss: 2.6166 (2.6166)  class_acc: 0.5898 (0.5898)  weight_decay: 0.0500 (0.0500)  time: 20.0462  data: 16.6315  max mem: 2928
Epoch: [279]  [200/625]  eta: 0:14:19  lr: 0.000055  min_lr: 0.000055  loss: 2.7202 (2.6871)  class_acc: 0.5898 (0.6010)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4478 (2.9410)  time: 1.9121  data: 0.6348  max mem: 2928
Epoch: [279]  [400/625]  eta: 0:07:24  lr: 0.000053  min_lr: 0.000053  loss: 2.7356 (2.6906)  class_acc: 0.5938 (0.6001)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.9114 (3.1683)  time: 2.0300  data: 0.0007  max mem: 2928
Epoch: [279]  [600/625]  eta: 0:00:48  lr: 0.000051  min_lr: 0.000051  loss: 2.6522 (2.6903)  class_acc: 0.5977 (0.5992)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3418 (3.1071)  time: 2.0551  data: 0.0293  max mem: 2928
Epoch: [279]  [624/625]  eta: 0:00:01  lr: 0.000051  min_lr: 0.000051  loss: 2.6741 (2.6903)  class_acc: 0.5859 (0.5991)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5262 (3.0888)  time: 0.7456  data: 0.0014  max mem: 2928
Epoch: [279] Total time: 0:19:48 (1.9011 s / it)
Averaged stats: lr: 0.000051  min_lr: 0.000051  loss: 2.6741 (2.6876)  class_acc: 0.5859 (0.5993)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5262 (3.0888)
Test:  [ 0/50]  eta: 0:10:15  loss: 1.3327 (1.3327)  acc1: 69.6000 (69.6000)  acc5: 90.4000 (90.4000)  time: 12.3071  data: 12.2686  max mem: 2928
Test:  [10/50]  eta: 0:01:22  loss: 1.2232 (1.2342)  acc1: 72.8000 (72.4364)  acc5: 91.2000 (90.3273)  time: 2.0663  data: 2.0443  max mem: 2928
Test:  [20/50]  eta: 0:00:48  loss: 1.4040 (1.3779)  acc1: 68.0000 (68.8000)  acc5: 89.6000 (88.8381)  time: 1.0731  data: 1.0531  max mem: 2928
Test:  [30/50]  eta: 0:00:26  loss: 1.5038 (1.3973)  acc1: 66.4000 (68.3097)  acc5: 87.2000 (88.3613)  time: 0.9030  data: 0.8841  max mem: 2928
Test:  [40/50]  eta: 0:00:10  loss: 1.3802 (1.4175)  acc1: 66.4000 (67.8244)  acc5: 87.2000 (87.8829)  time: 0.5052  data: 0.4863  max mem: 2928
Test:  [49/50]  eta: 0:00:00  loss: 1.4298 (1.4204)  acc1: 67.2000 (67.6480)  acc5: 87.2000 (87.7280)  time: 0.4532  data: 0.4344  max mem: 2928
Test: Total time: 0:00:44 (0.8894 s / it)
* Acc@1 68.532 Acc@5 88.082 loss 1.387
Accuracy of the model on the 50000 test images: 68.5%
Max accuracy: 68.53%
Epoch: [280]  [  0/625]  eta: 3:08:12  lr: 0.000051  min_lr: 0.000051  loss: 2.6742 (2.6742)  class_acc: 0.5859 (0.5859)  weight_decay: 0.0500 (0.0500)  time: 18.0682  data: 15.8064  max mem: 2928
Epoch: [280]  [200/625]  eta: 0:13:46  lr: 0.000050  min_lr: 0.000050  loss: 2.6397 (2.6768)  class_acc: 0.6094 (0.6026)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.2799 (inf)  time: 1.7707  data: 0.0006  max mem: 2928
Epoch: [280]  [400/625]  eta: 0:07:06  lr: 0.000048  min_lr: 0.000048  loss: 2.6548 (2.6809)  class_acc: 0.6055 (0.6001)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.2054 (inf)  time: 1.7875  data: 0.0007  max mem: 2928
Epoch: [280]  [600/625]  eta: 0:00:47  lr: 0.000046  min_lr: 0.000046  loss: 2.6748 (2.6851)  class_acc: 0.5938 (0.5990)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6952 (inf)  time: 1.9275  data: 0.0006  max mem: 2928
Epoch: [280]  [624/625]  eta: 0:00:01  lr: 0.000046  min_lr: 0.000046  loss: 2.6492 (2.6852)  class_acc: 0.6016 (0.5991)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5849 (inf)  time: 0.9609  data: 0.0014  max mem: 2928
Epoch: [280] Total time: 0:19:16 (1.8505 s / it)
Averaged stats: lr: 0.000046  min_lr: 0.000046  loss: 2.6492 (2.6893)  class_acc: 0.6016 (0.5988)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5849 (inf)
Test:  [ 0/50]  eta: 0:10:14  loss: 1.3498 (1.3498)  acc1: 67.2000 (67.2000)  acc5: 92.0000 (92.0000)  time: 12.2866  data: 12.2563  max mem: 2928
Test:  [10/50]  eta: 0:01:19  loss: 1.2368 (1.2383)  acc1: 72.8000 (72.5818)  acc5: 91.2000 (90.2546)  time: 1.9794  data: 1.9596  max mem: 2928
Test:  [20/50]  eta: 0:00:46  loss: 1.4004 (1.3801)  acc1: 68.8000 (69.4476)  acc5: 88.8000 (88.8000)  time: 1.0071  data: 0.9884  max mem: 2928
Test:  [30/50]  eta: 0:00:26  loss: 1.4768 (1.4000)  acc1: 66.4000 (68.6968)  acc5: 88.0000 (88.3871)  time: 0.9872  data: 0.9680  max mem: 2928
Test:  [40/50]  eta: 0:00:11  loss: 1.4075 (1.4208)  acc1: 66.4000 (68.1951)  acc5: 88.0000 (87.9610)  time: 0.6854  data: 0.6657  max mem: 2928
Test:  [49/50]  eta: 0:00:00  loss: 1.4243 (1.4253)  acc1: 65.6000 (67.9040)  acc5: 87.2000 (87.8080)  time: 0.5648  data: 0.5456  max mem: 2928
Test: Total time: 0:00:49 (0.9838 s / it)
* Acc@1 68.594 Acc@5 88.064 loss 1.391
Accuracy of the model on the 50000 test images: 68.6%
Max accuracy: 68.59%
Epoch: [281]  [  0/625]  eta: 3:13:44  lr: 0.000046  min_lr: 0.000046  loss: 2.8213 (2.8213)  class_acc: 0.5898 (0.5898)  weight_decay: 0.0500 (0.0500)  time: 18.5992  data: 18.4075  max mem: 2928
Epoch: [281]  [200/625]  eta: 0:13:50  lr: 0.000045  min_lr: 0.000045  loss: 2.7363 (2.6928)  class_acc: 0.5820 (0.5982)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7569 (2.9812)  time: 1.8773  data: 0.0008  max mem: 2928
Epoch: [281]  [400/625]  eta: 0:07:11  lr: 0.000043  min_lr: 0.000043  loss: 2.6870 (2.6889)  class_acc: 0.5898 (0.5991)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4591 (3.0920)  time: 2.0355  data: 0.0008  max mem: 2928
Epoch: [281]  [600/625]  eta: 0:00:48  lr: 0.000042  min_lr: 0.000042  loss: 2.6828 (2.6882)  class_acc: 0.6055 (0.5985)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3855 (3.0058)  time: 2.0052  data: 0.0008  max mem: 2928
Epoch: [281]  [624/625]  eta: 0:00:01  lr: 0.000042  min_lr: 0.000042  loss: 2.7301 (2.6895)  class_acc: 0.5977 (0.5985)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.9938 (3.0082)  time: 0.6779  data: 0.0020  max mem: 2928
Epoch: [281] Total time: 0:19:38 (1.8858 s / it)
Averaged stats: lr: 0.000042  min_lr: 0.000042  loss: 2.7301 (2.6852)  class_acc: 0.5977 (0.5997)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.9938 (3.0082)
Test:  [ 0/50]  eta: 0:10:12  loss: 1.3764 (1.3764)  acc1: 68.8000 (68.8000)  acc5: 90.4000 (90.4000)  time: 12.2577  data: 12.2286  max mem: 2928
Test:  [10/50]  eta: 0:01:29  loss: 1.2316 (1.2398)  acc1: 72.0000 (72.6545)  acc5: 90.4000 (90.1818)  time: 2.2326  data: 2.2137  max mem: 2928
Test:  [20/50]  eta: 0:00:54  loss: 1.3995 (1.3890)  acc1: 68.8000 (69.0667)  acc5: 88.8000 (88.6857)  time: 1.2890  data: 1.2701  max mem: 2928
Test:  [30/50]  eta: 0:00:31  loss: 1.5015 (1.4098)  acc1: 66.4000 (68.2581)  acc5: 88.0000 (88.0774)  time: 1.1875  data: 1.1672  max mem: 2928
Test:  [40/50]  eta: 0:00:12  loss: 1.4007 (1.4277)  acc1: 66.4000 (67.9024)  acc5: 87.2000 (87.7268)  time: 0.7521  data: 0.7328  max mem: 2928
Test:  [49/50]  eta: 0:00:01  loss: 1.4124 (1.4320)  acc1: 66.4000 (67.6800)  acc5: 87.2000 (87.5200)  time: 0.7510  data: 0.7327  max mem: 2928
Test: Total time: 0:00:53 (1.0675 s / it)
* Acc@1 68.550 Acc@5 88.018 loss 1.398
Accuracy of the model on the 50000 test images: 68.6%
Max accuracy: 68.59%
Epoch: [282]  [  0/625]  eta: 3:37:33  lr: 0.000042  min_lr: 0.000042  loss: 2.6832 (2.6832)  class_acc: 0.6211 (0.6211)  weight_decay: 0.0500 (0.0500)  time: 20.8848  data: 16.3276  max mem: 2928
Epoch: [282]  [200/625]  eta: 0:14:03  lr: 0.000040  min_lr: 0.000040  loss: 2.6873 (2.6818)  class_acc: 0.6055 (0.6021)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4936 (2.8106)  time: 1.7879  data: 0.0006  max mem: 2928
Epoch: [282]  [400/625]  eta: 0:07:13  lr: 0.000039  min_lr: 0.000039  loss: 2.6466 (2.6852)  class_acc: 0.6055 (0.6007)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6044 (3.0977)  time: 1.8668  data: 0.0150  max mem: 2928
Epoch: [282]  [600/625]  eta: 0:00:48  lr: 0.000037  min_lr: 0.000037  loss: 2.7311 (2.6832)  class_acc: 0.5898 (0.6019)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.8431 (3.1425)  time: 2.1169  data: 0.0180  max mem: 2928
Epoch: [282]  [624/625]  eta: 0:00:01  lr: 0.000037  min_lr: 0.000037  loss: 2.7014 (2.6830)  class_acc: 0.5898 (0.6019)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.0306 (3.1807)  time: 0.7332  data: 0.0193  max mem: 2928
Epoch: [282] Total time: 0:19:59 (1.9196 s / it)
Averaged stats: lr: 0.000037  min_lr: 0.000037  loss: 2.7014 (2.6850)  class_acc: 0.5898 (0.5998)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.0306 (3.1807)
Test:  [ 0/50]  eta: 0:10:20  loss: 1.3544 (1.3544)  acc1: 67.2000 (67.2000)  acc5: 90.4000 (90.4000)  time: 12.4064  data: 12.3627  max mem: 2928
Test:  [10/50]  eta: 0:01:21  loss: 1.2201 (1.2334)  acc1: 73.6000 (72.8727)  acc5: 91.2000 (90.3273)  time: 2.0299  data: 2.0079  max mem: 2928
Test:  [20/50]  eta: 0:00:47  loss: 1.3980 (1.3812)  acc1: 69.6000 (69.4095)  acc5: 89.6000 (88.8762)  time: 1.0432  data: 1.0233  max mem: 2928
Test:  [30/50]  eta: 0:00:27  loss: 1.5038 (1.4016)  acc1: 66.4000 (68.6968)  acc5: 88.0000 (88.4645)  time: 1.0084  data: 0.9883  max mem: 2928
Test:  [40/50]  eta: 0:00:11  loss: 1.4008 (1.4199)  acc1: 67.2000 (68.0390)  acc5: 87.2000 (87.9610)  time: 0.6561  data: 0.6362  max mem: 2928
Test:  [49/50]  eta: 0:00:00  loss: 1.4060 (1.4226)  acc1: 65.6000 (67.8240)  acc5: 87.2000 (87.8080)  time: 0.6545  data: 0.6361  max mem: 2928
Test: Total time: 0:00:46 (0.9336 s / it)
* Acc@1 68.676 Acc@5 88.124 loss 1.389
Accuracy of the model on the 50000 test images: 68.7%
Max accuracy: 68.68%
Epoch: [283]  [  0/625]  eta: 3:30:01  lr: 0.000037  min_lr: 0.000037  loss: 2.5470 (2.5470)  class_acc: 0.6680 (0.6680)  weight_decay: 0.0500 (0.0500)  time: 20.1618  data: 17.7754  max mem: 2928
Epoch: [283]  [200/625]  eta: 0:14:36  lr: 0.000036  min_lr: 0.000036  loss: 2.6979 (2.6966)  class_acc: 0.6094 (0.5982)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3781 (2.9006)  time: 1.7888  data: 0.0010  max mem: 2928
Epoch: [283]  [400/625]  eta: 0:07:29  lr: 0.000035  min_lr: 0.000035  loss: 2.6904 (2.6868)  class_acc: 0.5977 (0.5998)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5846 (2.9859)  time: 1.9854  data: 0.0008  max mem: 2928
Epoch: [283]  [600/625]  eta: 0:00:49  lr: 0.000033  min_lr: 0.000033  loss: 2.6247 (2.6860)  class_acc: 0.5938 (0.6000)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6765 (2.9870)  time: 1.9493  data: 0.0019  max mem: 2928
Epoch: [283]  [624/625]  eta: 0:00:01  lr: 0.000033  min_lr: 0.000033  loss: 2.6755 (2.6857)  class_acc: 0.5938 (0.5999)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7766 (2.9764)  time: 0.5867  data: 0.0023  max mem: 2928
Epoch: [283] Total time: 0:20:31 (1.9708 s / it)
Averaged stats: lr: 0.000033  min_lr: 0.000033  loss: 2.6755 (2.6852)  class_acc: 0.5938 (0.5996)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7766 (2.9764)
Test:  [ 0/50]  eta: 0:10:44  loss: 1.3511 (1.3511)  acc1: 67.2000 (67.2000)  acc5: 91.2000 (91.2000)  time: 12.8905  data: 12.8601  max mem: 2928
Test:  [10/50]  eta: 0:01:22  loss: 1.2415 (1.2416)  acc1: 72.8000 (73.0182)  acc5: 91.2000 (90.3273)  time: 2.0600  data: 2.0403  max mem: 2928
Test:  [20/50]  eta: 0:00:47  loss: 1.3866 (1.3816)  acc1: 68.8000 (69.1429)  acc5: 88.8000 (88.9143)  time: 1.0117  data: 0.9918  max mem: 2928
Test:  [30/50]  eta: 0:00:27  loss: 1.4887 (1.4011)  acc1: 66.4000 (68.6194)  acc5: 88.0000 (88.3355)  time: 1.0095  data: 0.9902  max mem: 2928
Test:  [40/50]  eta: 0:00:11  loss: 1.4002 (1.4201)  acc1: 66.4000 (68.1171)  acc5: 87.2000 (87.9610)  time: 0.7426  data: 0.7232  max mem: 2928
Test:  [49/50]  eta: 0:00:01  loss: 1.4248 (1.4233)  acc1: 65.6000 (67.7120)  acc5: 87.2000 (87.7600)  time: 0.7130  data: 0.6932  max mem: 2928
Test: Total time: 0:00:50 (1.0134 s / it)
* Acc@1 68.714 Acc@5 88.108 loss 1.392
Accuracy of the model on the 50000 test images: 68.7%
Max accuracy: 68.71%
Epoch: [284]  [  0/625]  eta: 3:34:22  lr: 0.000033  min_lr: 0.000033  loss: 2.7053 (2.7053)  class_acc: 0.5742 (0.5742)  weight_decay: 0.0500 (0.0500)  time: 20.5796  data: 17.8878  max mem: 2928
Epoch: [284]  [200/625]  eta: 0:14:12  lr: 0.000032  min_lr: 0.000032  loss: 2.6803 (2.6917)  class_acc: 0.5938 (0.5967)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5237 (3.0243)  time: 1.8779  data: 0.0005  max mem: 2928
Epoch: [284]  [400/625]  eta: 0:07:11  lr: 0.000031  min_lr: 0.000031  loss: 2.6899 (2.6887)  class_acc: 0.5938 (0.5986)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7147 (3.0954)  time: 1.7720  data: 0.0006  max mem: 2928
Epoch: [284]  [600/625]  eta: 0:00:47  lr: 0.000029  min_lr: 0.000029  loss: 2.7134 (2.6882)  class_acc: 0.5781 (0.5985)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6319 (3.0402)  time: 2.0573  data: 0.0008  max mem: 2928
Epoch: [284]  [624/625]  eta: 0:00:01  lr: 0.000029  min_lr: 0.000029  loss: 2.7230 (2.6890)  class_acc: 0.5781 (0.5982)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.8797 (3.0550)  time: 0.7659  data: 0.0012  max mem: 2928
Epoch: [284] Total time: 0:19:44 (1.8960 s / it)
Averaged stats: lr: 0.000029  min_lr: 0.000029  loss: 2.7230 (2.6841)  class_acc: 0.5781 (0.6003)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.8797 (3.0550)
Test:  [ 0/50]  eta: 0:10:30  loss: 1.3430 (1.3430)  acc1: 65.6000 (65.6000)  acc5: 92.0000 (92.0000)  time: 12.6097  data: 12.5836  max mem: 2928
Test:  [10/50]  eta: 0:01:22  loss: 1.2264 (1.2395)  acc1: 74.4000 (72.8727)  acc5: 91.2000 (90.4000)  time: 2.0679  data: 2.0481  max mem: 2928
Test:  [20/50]  eta: 0:00:49  loss: 1.4008 (1.3874)  acc1: 68.8000 (69.2952)  acc5: 89.6000 (88.9143)  time: 1.0865  data: 1.0676  max mem: 2928
Test:  [30/50]  eta: 0:00:28  loss: 1.5308 (1.4051)  acc1: 66.4000 (68.6194)  acc5: 87.2000 (88.4903)  time: 1.0803  data: 1.0617  max mem: 2928
Test:  [40/50]  eta: 0:00:11  loss: 1.4155 (1.4214)  acc1: 66.4000 (68.0585)  acc5: 88.0000 (88.0000)  time: 0.7159  data: 0.6965  max mem: 2928
Test:  [49/50]  eta: 0:00:00  loss: 1.3998 (1.4254)  acc1: 65.6000 (67.8400)  acc5: 87.2000 (87.7440)  time: 0.6736  data: 0.6530  max mem: 2928
Test: Total time: 0:00:49 (0.9871 s / it)
* Acc@1 68.628 Acc@5 88.056 loss 1.392
Accuracy of the model on the 50000 test images: 68.6%
Max accuracy: 68.71%
Epoch: [285]  [  0/625]  eta: 3:37:54  lr: 0.000029  min_lr: 0.000029  loss: 2.6885 (2.6885)  class_acc: 0.6484 (0.6484)  weight_decay: 0.0500 (0.0500)  time: 20.9187  data: 17.2003  max mem: 2928
Epoch: [285]  [200/625]  eta: 0:14:08  lr: 0.000028  min_lr: 0.000028  loss: 2.6788 (2.6842)  class_acc: 0.6016 (0.5993)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3959 (2.7854)  time: 1.9613  data: 0.0010  max mem: 2928
Epoch: [285]  [400/625]  eta: 0:07:19  lr: 0.000027  min_lr: 0.000027  loss: 2.6458 (2.6777)  class_acc: 0.5977 (0.6010)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6664 (2.9210)  time: 2.0289  data: 0.0008  max mem: 2928
Epoch: [285]  [600/625]  eta: 0:00:49  lr: 0.000026  min_lr: 0.000026  loss: 2.6542 (2.6781)  class_acc: 0.5977 (0.6007)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6703 (2.9270)  time: 2.1352  data: 0.0014  max mem: 2928
Epoch: [285]  [624/625]  eta: 0:00:01  lr: 0.000026  min_lr: 0.000026  loss: 2.6575 (2.6782)  class_acc: 0.5938 (0.6005)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4929 (2.9171)  time: 0.7751  data: 0.0018  max mem: 2928
Epoch: [285] Total time: 0:20:04 (1.9270 s / it)
Averaged stats: lr: 0.000026  min_lr: 0.000026  loss: 2.6575 (2.6807)  class_acc: 0.5938 (0.6005)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4929 (2.9171)
Test:  [ 0/50]  eta: 0:10:50  loss: 1.3308 (1.3308)  acc1: 66.4000 (66.4000)  acc5: 91.2000 (91.2000)  time: 13.0177  data: 12.9864  max mem: 2928
Test:  [10/50]  eta: 0:01:25  loss: 1.2131 (1.2304)  acc1: 73.6000 (73.7455)  acc5: 91.2000 (90.6182)  time: 2.1352  data: 2.1147  max mem: 2928
Test:  [20/50]  eta: 0:00:50  loss: 1.4042 (1.3776)  acc1: 68.8000 (69.7524)  acc5: 88.8000 (89.0667)  time: 1.1125  data: 1.0941  max mem: 2928
Test:  [30/50]  eta: 0:00:30  loss: 1.4835 (1.3992)  acc1: 67.2000 (69.0065)  acc5: 88.8000 (88.5677)  time: 1.1653  data: 1.1471  max mem: 2928
Test:  [40/50]  eta: 0:00:13  loss: 1.3978 (1.4179)  acc1: 66.4000 (68.3512)  acc5: 88.0000 (88.0390)  time: 1.0266  data: 1.0079  max mem: 2928
Test:  [49/50]  eta: 0:00:01  loss: 1.4277 (1.4210)  acc1: 67.2000 (68.0640)  acc5: 87.2000 (87.8400)  time: 0.9427  data: 0.9205  max mem: 2928
Test: Total time: 0:00:56 (1.1364 s / it)
* Acc@1 68.716 Acc@5 88.110 loss 1.389
Accuracy of the model on the 50000 test images: 68.7%
Max accuracy: 68.72%
Epoch: [286]  [  0/625]  eta: 3:26:44  lr: 0.000026  min_lr: 0.000026  loss: 2.8813 (2.8813)  class_acc: 0.5742 (0.5742)  weight_decay: 0.0500 (0.0500)  time: 19.8472  data: 18.6266  max mem: 2928
Epoch: [286]  [200/625]  eta: 0:14:02  lr: 0.000025  min_lr: 0.000025  loss: 2.6691 (2.6869)  class_acc: 0.5977 (0.5998)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.2221 (3.1651)  time: 1.8964  data: 0.0006  max mem: 2928
Epoch: [286]  [400/625]  eta: 0:07:23  lr: 0.000023  min_lr: 0.000023  loss: 2.6892 (2.6783)  class_acc: 0.5977 (0.6016)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3608 (3.0207)  time: 1.8217  data: 0.0319  max mem: 2928
Epoch: [286]  [600/625]  eta: 0:00:48  lr: 0.000022  min_lr: 0.000022  loss: 2.7525 (2.6787)  class_acc: 0.5820 (0.6025)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6531 (inf)  time: 1.9520  data: 0.1765  max mem: 2928
Epoch: [286]  [624/625]  eta: 0:00:01  lr: 0.000022  min_lr: 0.000022  loss: 2.6599 (2.6794)  class_acc: 0.6133 (0.6026)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6172 (inf)  time: 0.9064  data: 0.1038  max mem: 2928
Epoch: [286] Total time: 0:19:57 (1.9163 s / it)
Averaged stats: lr: 0.000022  min_lr: 0.000022  loss: 2.6599 (2.6829)  class_acc: 0.6133 (0.6003)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6172 (inf)
Test:  [ 0/50]  eta: 0:10:05  loss: 1.3468 (1.3468)  acc1: 64.0000 (64.0000)  acc5: 90.4000 (90.4000)  time: 12.1065  data: 12.0783  max mem: 2928
Test:  [10/50]  eta: 0:01:21  loss: 1.2118 (1.2286)  acc1: 73.6000 (73.2364)  acc5: 91.2000 (90.3273)  time: 2.0423  data: 2.0221  max mem: 2928
Test:  [20/50]  eta: 0:00:48  loss: 1.3857 (1.3784)  acc1: 69.6000 (69.6381)  acc5: 88.0000 (88.8000)  time: 1.0810  data: 1.0599  max mem: 2928
Test:  [30/50]  eta: 0:00:27  loss: 1.5108 (1.3973)  acc1: 66.4000 (68.8258)  acc5: 87.2000 (88.3613)  time: 1.0048  data: 0.9822  max mem: 2928
Test:  [40/50]  eta: 0:00:11  loss: 1.3972 (1.4171)  acc1: 66.4000 (68.1561)  acc5: 88.0000 (87.8244)  time: 0.6052  data: 0.5806  max mem: 2928
Test:  [49/50]  eta: 0:00:00  loss: 1.4134 (1.4198)  acc1: 67.2000 (67.9680)  acc5: 87.2000 (87.6640)  time: 0.5323  data: 0.5102  max mem: 2928
Test: Total time: 0:00:46 (0.9324 s / it)
* Acc@1 68.756 Acc@5 88.172 loss 1.385
Accuracy of the model on the 50000 test images: 68.8%
Max accuracy: 68.76%
Epoch: [287]  [  0/625]  eta: 3:20:48  lr: 0.000022  min_lr: 0.000022  loss: 2.5877 (2.5877)  class_acc: 0.6406 (0.6406)  weight_decay: 0.0500 (0.0500)  time: 19.2779  data: 15.3315  max mem: 2928
Epoch: [287]  [200/625]  eta: 0:13:59  lr: 0.000021  min_lr: 0.000021  loss: 2.6888 (2.6895)  class_acc: 0.5938 (0.6009)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6288 (2.8720)  time: 1.8534  data: 0.0007  max mem: 2928
Epoch: [287]  [400/625]  eta: 0:07:19  lr: 0.000020  min_lr: 0.000020  loss: 2.6429 (2.6867)  class_acc: 0.6016 (0.6020)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7867 (2.8569)  time: 2.0474  data: 0.0007  max mem: 2928
Epoch: [287]  [600/625]  eta: 0:00:49  lr: 0.000019  min_lr: 0.000019  loss: 2.6753 (2.6856)  class_acc: 0.5938 (0.6017)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.0845 (2.9335)  time: 1.8874  data: 0.0007  max mem: 2928
Epoch: [287]  [624/625]  eta: 0:00:01  lr: 0.000019  min_lr: 0.000019  loss: 2.6694 (2.6847)  class_acc: 0.6094 (0.6020)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2659 (2.9257)  time: 0.8055  data: 0.0014  max mem: 2928
Epoch: [287] Total time: 0:20:00 (1.9207 s / it)
Averaged stats: lr: 0.000019  min_lr: 0.000019  loss: 2.6694 (2.6800)  class_acc: 0.6094 (0.6011)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2659 (2.9257)
Test:  [ 0/50]  eta: 0:09:49  loss: 1.3451 (1.3451)  acc1: 66.4000 (66.4000)  acc5: 91.2000 (91.2000)  time: 11.7954  data: 11.7609  max mem: 2928
Test:  [10/50]  eta: 0:01:21  loss: 1.2081 (1.2299)  acc1: 72.8000 (73.0182)  acc5: 90.4000 (90.2545)  time: 2.0269  data: 2.0067  max mem: 2928
Test:  [20/50]  eta: 0:00:48  loss: 1.3801 (1.3794)  acc1: 68.8000 (69.5238)  acc5: 88.8000 (89.0286)  time: 1.1215  data: 1.0998  max mem: 2928
Test:  [30/50]  eta: 0:00:29  loss: 1.5131 (1.3977)  acc1: 67.2000 (69.0323)  acc5: 88.8000 (88.4645)  time: 1.1378  data: 1.1160  max mem: 2928
Test:  [40/50]  eta: 0:00:13  loss: 1.3969 (1.4160)  acc1: 67.2000 (68.3317)  acc5: 88.0000 (88.0195)  time: 1.0370  data: 1.0182  max mem: 2928
Test:  [49/50]  eta: 0:00:01  loss: 1.4230 (1.4204)  acc1: 66.4000 (67.9680)  acc5: 86.4000 (87.7760)  time: 0.9709  data: 0.9500  max mem: 2928
Test: Total time: 0:00:56 (1.1316 s / it)
* Acc@1 68.722 Acc@5 88.126 loss 1.386
Accuracy of the model on the 50000 test images: 68.7%
Max accuracy: 68.76%
Epoch: [288]  [  0/625]  eta: 3:52:42  lr: 0.000019  min_lr: 0.000019  loss: 2.5230 (2.5230)  class_acc: 0.6406 (0.6406)  weight_decay: 0.0500 (0.0500)  time: 22.3404  data: 18.7238  max mem: 2928
Epoch: [288]  [200/625]  eta: 0:14:18  lr: 0.000018  min_lr: 0.000018  loss: 2.6787 (2.6751)  class_acc: 0.5938 (0.6003)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5655 (2.8107)  time: 1.8712  data: 0.0005  max mem: 2928
Epoch: [288]  [400/625]  eta: 0:07:24  lr: 0.000017  min_lr: 0.000017  loss: 2.7081 (2.6745)  class_acc: 0.6016 (0.6017)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.8706 (2.8939)  time: 1.7816  data: 0.0005  max mem: 2928
Epoch: [288]  [600/625]  eta: 0:00:49  lr: 0.000016  min_lr: 0.000016  loss: 2.6662 (2.6764)  class_acc: 0.5977 (0.6011)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7887 (2.8693)  time: 2.0209  data: 0.0006  max mem: 2928
Epoch: [288]  [624/625]  eta: 0:00:01  lr: 0.000016  min_lr: 0.000016  loss: 2.7278 (2.6776)  class_acc: 0.5898 (0.6007)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4043 (2.8600)  time: 0.7411  data: 0.0018  max mem: 2928
Epoch: [288] Total time: 0:20:07 (1.9319 s / it)
Averaged stats: lr: 0.000016  min_lr: 0.000016  loss: 2.7278 (2.6808)  class_acc: 0.5898 (0.6011)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4043 (2.8600)
Test:  [ 0/50]  eta: 0:10:05  loss: 1.3328 (1.3328)  acc1: 66.4000 (66.4000)  acc5: 91.2000 (91.2000)  time: 12.1174  data: 12.0820  max mem: 2928
Test:  [10/50]  eta: 0:01:21  loss: 1.2014 (1.2246)  acc1: 73.6000 (73.3818)  acc5: 91.2000 (90.4727)  time: 2.0342  data: 2.0128  max mem: 2928
Test:  [20/50]  eta: 0:00:47  loss: 1.3817 (1.3747)  acc1: 68.8000 (69.3714)  acc5: 89.6000 (88.9905)  time: 1.0731  data: 1.0534  max mem: 2928
Test:  [30/50]  eta: 0:00:27  loss: 1.4977 (1.3945)  acc1: 67.2000 (68.8516)  acc5: 88.0000 (88.5161)  time: 1.0252  data: 1.0042  max mem: 2928
Test:  [40/50]  eta: 0:00:11  loss: 1.3696 (1.4138)  acc1: 67.2000 (68.1951)  acc5: 88.0000 (88.0195)  time: 0.6460  data: 0.6245  max mem: 2928
Test:  [49/50]  eta: 0:00:00  loss: 1.4113 (1.4168)  acc1: 67.2000 (67.9040)  acc5: 87.2000 (87.8720)  time: 0.5907  data: 0.5707  max mem: 2928
Test: Total time: 0:00:46 (0.9374 s / it)
* Acc@1 68.806 Acc@5 88.178 loss 1.385
Accuracy of the model on the 50000 test images: 68.8%
Max accuracy: 68.81%
Epoch: [289]  [  0/625]  eta: 3:26:39  lr: 0.000016  min_lr: 0.000016  loss: 2.8413 (2.8413)  class_acc: 0.5703 (0.5703)  weight_decay: 0.0500 (0.0500)  time: 19.8397  data: 19.2889  max mem: 2928
Epoch: [289]  [200/625]  eta: 0:14:42  lr: 0.000015  min_lr: 0.000015  loss: 2.6746 (2.6934)  class_acc: 0.6016 (0.5988)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4621 (2.8588)  time: 1.9523  data: 0.0178  max mem: 2928
Epoch: [289]  [400/625]  eta: 0:07:31  lr: 0.000014  min_lr: 0.000014  loss: 2.6586 (2.6925)  class_acc: 0.6055 (0.5993)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5300 (2.8845)  time: 1.9432  data: 0.0043  max mem: 2928
Epoch: [289]  [600/625]  eta: 0:00:49  lr: 0.000014  min_lr: 0.000014  loss: 2.6777 (2.6879)  class_acc: 0.5898 (0.5999)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4942 (2.8763)  time: 1.9701  data: 0.0007  max mem: 2928
Epoch: [289]  [624/625]  eta: 0:00:01  lr: 0.000014  min_lr: 0.000014  loss: 2.6521 (2.6882)  class_acc: 0.5977 (0.5996)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7526 (2.8755)  time: 0.7331  data: 0.0120  max mem: 2928
Epoch: [289] Total time: 0:20:11 (1.9389 s / it)
Averaged stats: lr: 0.000014  min_lr: 0.000014  loss: 2.6521 (2.6813)  class_acc: 0.5977 (0.6010)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7526 (2.8755)
Test:  [ 0/50]  eta: 0:09:09  loss: 1.3287 (1.3287)  acc1: 66.4000 (66.4000)  acc5: 92.0000 (92.0000)  time: 10.9822  data: 10.9572  max mem: 2928
Test:  [10/50]  eta: 0:01:13  loss: 1.2106 (1.2306)  acc1: 73.6000 (73.3818)  acc5: 91.2000 (90.6182)  time: 1.8441  data: 1.8242  max mem: 2928
Test:  [20/50]  eta: 0:00:43  loss: 1.3901 (1.3761)  acc1: 69.6000 (69.7905)  acc5: 89.6000 (89.2191)  time: 0.9655  data: 0.9463  max mem: 2928
Test:  [30/50]  eta: 0:00:25  loss: 1.5018 (1.3950)  acc1: 66.4000 (68.9806)  acc5: 88.8000 (88.6968)  time: 0.9590  data: 0.9380  max mem: 2928
Test:  [40/50]  eta: 0:00:11  loss: 1.3728 (1.4133)  acc1: 66.4000 (68.2732)  acc5: 87.2000 (88.1366)  time: 0.8484  data: 0.8241  max mem: 2928
Test:  [49/50]  eta: 0:00:01  loss: 1.4242 (1.4170)  acc1: 67.2000 (67.9680)  acc5: 87.2000 (87.9520)  time: 0.6265  data: 0.6049  max mem: 2928
Test: Total time: 0:00:50 (1.0103 s / it)
* Acc@1 68.856 Acc@5 88.246 loss 1.382
Accuracy of the model on the 50000 test images: 68.9%
Max accuracy: 68.86%
Epoch: [290]  [  0/625]  eta: 3:40:22  lr: 0.000014  min_lr: 0.000014  loss: 2.7858 (2.7858)  class_acc: 0.5625 (0.5625)  weight_decay: 0.0500 (0.0500)  time: 21.1559  data: 17.8987  max mem: 2928
Epoch: [290]  [200/625]  eta: 0:14:25  lr: 0.000013  min_lr: 0.000013  loss: 2.6827 (2.6674)  class_acc: 0.6094 (0.6013)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6508 (2.8644)  time: 2.0875  data: 0.0151  max mem: 2928
Epoch: [290]  [400/625]  eta: 0:07:26  lr: 0.000012  min_lr: 0.000012  loss: 2.6567 (2.6753)  class_acc: 0.6016 (0.6016)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.8170 (2.8953)  time: 1.9420  data: 0.0007  max mem: 2928
Epoch: [290]  [600/625]  eta: 0:00:49  lr: 0.000011  min_lr: 0.000011  loss: 2.6216 (2.6785)  class_acc: 0.5977 (0.6004)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3132 (2.8436)  time: 2.0309  data: 0.0008  max mem: 2928
Epoch: [290]  [624/625]  eta: 0:00:01  lr: 0.000011  min_lr: 0.000011  loss: 2.6949 (2.6797)  class_acc: 0.6016 (0.6002)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6061 (2.8500)  time: 0.7422  data: 0.0020  max mem: 2928
Epoch: [290] Total time: 0:20:05 (1.9286 s / it)
Averaged stats: lr: 0.000011  min_lr: 0.000011  loss: 2.6949 (2.6799)  class_acc: 0.6016 (0.6009)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6061 (2.8500)
Test:  [ 0/50]  eta: 0:08:30  loss: 1.3376 (1.3376)  acc1: 65.6000 (65.6000)  acc5: 91.2000 (91.2000)  time: 10.2041  data: 10.1787  max mem: 2928
Test:  [10/50]  eta: 0:01:15  loss: 1.2093 (1.2255)  acc1: 72.8000 (73.5273)  acc5: 91.2000 (90.5455)  time: 1.8853  data: 1.8659  max mem: 2928
Test:  [20/50]  eta: 0:00:43  loss: 1.3743 (1.3709)  acc1: 69.6000 (70.0191)  acc5: 89.6000 (89.2191)  time: 1.0109  data: 0.9911  max mem: 2928
Test:  [30/50]  eta: 0:00:25  loss: 1.4948 (1.3915)  acc1: 67.2000 (69.1613)  acc5: 88.8000 (88.7484)  time: 0.9406  data: 0.9205  max mem: 2928
Test:  [40/50]  eta: 0:00:11  loss: 1.3753 (1.4111)  acc1: 65.6000 (68.4293)  acc5: 88.0000 (88.1951)  time: 0.8308  data: 0.8118  max mem: 2928
Test:  [49/50]  eta: 0:00:01  loss: 1.4205 (1.4145)  acc1: 66.4000 (68.1600)  acc5: 87.2000 (87.9680)  time: 0.7311  data: 0.7104  max mem: 2928
Test: Total time: 0:00:50 (1.0103 s / it)
* Acc@1 68.848 Acc@5 88.184 loss 1.381
Accuracy of the model on the 50000 test images: 68.8%
Max accuracy: 68.86%
Epoch: [291]  [  0/625]  eta: 3:42:43  lr: 0.000011  min_lr: 0.000011  loss: 2.8587 (2.8587)  class_acc: 0.5469 (0.5469)  weight_decay: 0.0500 (0.0500)  time: 21.3814  data: 17.1549  max mem: 2928
Epoch: [291]  [200/625]  eta: 0:13:48  lr: 0.000010  min_lr: 0.000010  loss: 2.7152 (2.6814)  class_acc: 0.5938 (0.5983)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3529 (2.8354)  time: 1.8143  data: 0.0006  max mem: 2928
Epoch: [291]  [400/625]  eta: 0:07:14  lr: 0.000010  min_lr: 0.000010  loss: 2.6895 (2.6801)  class_acc: 0.5977 (0.5992)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4969 (2.8554)  time: 1.9893  data: 0.0006  max mem: 2928
Epoch: [291]  [600/625]  eta: 0:00:48  lr: 0.000009  min_lr: 0.000009  loss: 2.6592 (2.6761)  class_acc: 0.6016 (0.6012)  weight_decay: 0.0500 (0.0500)  grad_norm: 3.0971 (2.9058)  time: 1.9543  data: 0.0007  max mem: 2928
Epoch: [291]  [624/625]  eta: 0:00:01  lr: 0.000009  min_lr: 0.000009  loss: 2.6887 (2.6766)  class_acc: 0.5977 (0.6012)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7518 (2.8979)  time: 0.7669  data: 0.0016  max mem: 2928
Epoch: [291] Total time: 0:20:08 (1.9331 s / it)
Averaged stats: lr: 0.000009  min_lr: 0.000009  loss: 2.6887 (2.6786)  class_acc: 0.5977 (0.6013)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7518 (2.8979)
Test:  [ 0/50]  eta: 0:10:01  loss: 1.3369 (1.3369)  acc1: 67.2000 (67.2000)  acc5: 91.2000 (91.2000)  time: 12.0279  data: 11.9919  max mem: 2928
Test:  [10/50]  eta: 0:01:20  loss: 1.2133 (1.2291)  acc1: 72.8000 (73.3091)  acc5: 91.2000 (90.2546)  time: 2.0021  data: 1.9807  max mem: 2928
Test:  [20/50]  eta: 0:00:47  loss: 1.3838 (1.3757)  acc1: 68.8000 (69.6000)  acc5: 88.8000 (89.0667)  time: 1.0465  data: 1.0272  max mem: 2928
Test:  [30/50]  eta: 0:00:28  loss: 1.5077 (1.3953)  acc1: 66.4000 (69.0065)  acc5: 88.8000 (88.6194)  time: 1.0694  data: 1.0496  max mem: 2928
Test:  [40/50]  eta: 0:00:12  loss: 1.3790 (1.4147)  acc1: 66.4000 (68.2927)  acc5: 87.2000 (88.1171)  time: 0.9021  data: 0.8814  max mem: 2928
Test:  [49/50]  eta: 0:00:01  loss: 1.4154 (1.4183)  acc1: 66.4000 (68.0480)  acc5: 87.2000 (87.9520)  time: 0.8812  data: 0.8615  max mem: 2928
Test: Total time: 0:00:55 (1.1099 s / it)
* Acc@1 68.828 Acc@5 88.200 loss 1.384
Accuracy of the model on the 50000 test images: 68.8%
Max accuracy: 68.86%
Epoch: [292]  [  0/625]  eta: 3:25:38  lr: 0.000009  min_lr: 0.000009  loss: 2.6794 (2.6794)  class_acc: 0.5625 (0.5625)  weight_decay: 0.0500 (0.0500)  time: 19.7421  data: 19.6194  max mem: 2928
Epoch: [292]  [200/625]  eta: 0:14:35  lr: 0.000008  min_lr: 0.000008  loss: 2.6573 (2.6768)  class_acc: 0.6094 (0.6014)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4938 (2.6780)  time: 1.8099  data: 0.0156  max mem: 2928
Epoch: [292]  [400/625]  eta: 0:07:24  lr: 0.000008  min_lr: 0.000008  loss: 2.6593 (2.6772)  class_acc: 0.5938 (0.6006)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7928 (2.7951)  time: 2.0499  data: 0.0006  max mem: 2928
Epoch: [292]  [600/625]  eta: 0:00:49  lr: 0.000007  min_lr: 0.000007  loss: 2.6311 (2.6752)  class_acc: 0.5898 (0.6014)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4127 (2.7912)  time: 2.1412  data: 0.0005  max mem: 2928
Epoch: [292]  [624/625]  eta: 0:00:01  lr: 0.000007  min_lr: 0.000007  loss: 2.6551 (2.6758)  class_acc: 0.6016 (0.6013)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2329 (2.7846)  time: 0.7125  data: 0.0013  max mem: 2928
Epoch: [292] Total time: 0:20:08 (1.9339 s / it)
Averaged stats: lr: 0.000007  min_lr: 0.000007  loss: 2.6551 (2.6780)  class_acc: 0.6016 (0.6016)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2329 (2.7846)
Test:  [ 0/50]  eta: 0:09:29  loss: 1.3263 (1.3263)  acc1: 65.6000 (65.6000)  acc5: 91.2000 (91.2000)  time: 11.3945  data: 11.3689  max mem: 2928
Test:  [10/50]  eta: 0:01:19  loss: 1.2212 (1.2353)  acc1: 73.6000 (73.0909)  acc5: 91.2000 (90.3273)  time: 1.9804  data: 1.9588  max mem: 2928
Test:  [20/50]  eta: 0:00:47  loss: 1.4004 (1.3811)  acc1: 68.8000 (69.5238)  acc5: 89.6000 (89.0667)  time: 1.0940  data: 1.0726  max mem: 2928
Test:  [30/50]  eta: 0:00:27  loss: 1.4957 (1.3997)  acc1: 68.0000 (68.9290)  acc5: 88.8000 (88.5419)  time: 1.0265  data: 1.0051  max mem: 2928
Test:  [40/50]  eta: 0:00:11  loss: 1.3843 (1.4179)  acc1: 67.2000 (68.2927)  acc5: 88.0000 (88.0976)  time: 0.6181  data: 0.5965  max mem: 2928
Test:  [49/50]  eta: 0:00:00  loss: 1.4216 (1.4216)  acc1: 65.6000 (67.9200)  acc5: 87.2000 (87.9200)  time: 0.5584  data: 0.5383  max mem: 2928
Test: Total time: 0:00:46 (0.9350 s / it)
* Acc@1 68.770 Acc@5 88.210 loss 1.389
Accuracy of the model on the 50000 test images: 68.8%
Max accuracy: 68.86%
Epoch: [293]  [  0/625]  eta: 3:54:21  lr: 0.000007  min_lr: 0.000007  loss: 2.7472 (2.7472)  class_acc: 0.6328 (0.6328)  weight_decay: 0.0500 (0.0500)  time: 22.4990  data: 16.0056  max mem: 2928
Epoch: [293]  [200/625]  eta: 0:14:25  lr: 0.000007  min_lr: 0.000007  loss: 2.6708 (2.6846)  class_acc: 0.6094 (0.6008)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5527 (2.8190)  time: 1.8553  data: 0.0007  max mem: 2928
Epoch: [293]  [400/625]  eta: 0:07:25  lr: 0.000006  min_lr: 0.000006  loss: 2.6712 (2.6860)  class_acc: 0.6211 (0.6007)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6632 (inf)  time: 1.9945  data: 0.0006  max mem: 2928
Epoch: [293]  [600/625]  eta: 0:00:49  lr: 0.000006  min_lr: 0.000006  loss: 2.7119 (2.6856)  class_acc: 0.5820 (0.6008)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.1901 (inf)  time: 2.0840  data: 0.0007  max mem: 2928
Epoch: [293]  [624/625]  eta: 0:00:01  lr: 0.000006  min_lr: 0.000006  loss: 2.6658 (2.6851)  class_acc: 0.5898 (0.6007)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2412 (inf)  time: 0.7882  data: 0.0016  max mem: 2928
Epoch: [293] Total time: 0:20:03 (1.9263 s / it)
Averaged stats: lr: 0.000006  min_lr: 0.000006  loss: 2.6658 (2.6794)  class_acc: 0.5898 (0.6016)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.2412 (inf)
Test:  [ 0/50]  eta: 0:10:36  loss: 1.3297 (1.3297)  acc1: 65.6000 (65.6000)  acc5: 91.2000 (91.2000)  time: 12.7391  data: 12.7130  max mem: 2928
Test:  [10/50]  eta: 0:01:28  loss: 1.2089 (1.2266)  acc1: 72.8000 (73.0909)  acc5: 91.2000 (90.4000)  time: 2.2219  data: 2.2026  max mem: 2928
Test:  [20/50]  eta: 0:00:53  loss: 1.3835 (1.3735)  acc1: 68.0000 (69.3714)  acc5: 88.8000 (89.1429)  time: 1.2317  data: 1.2132  max mem: 2928
Test:  [30/50]  eta: 0:00:32  loss: 1.4962 (1.3934)  acc1: 67.2000 (68.7484)  acc5: 88.8000 (88.6194)  time: 1.2789  data: 1.2600  max mem: 2928
Test:  [40/50]  eta: 0:00:13  loss: 1.3787 (1.4113)  acc1: 67.2000 (68.1561)  acc5: 88.0000 (88.1951)  time: 0.9704  data: 0.9517  max mem: 2928
Test:  [49/50]  eta: 0:00:01  loss: 1.4079 (1.4145)  acc1: 67.2000 (67.9840)  acc5: 87.2000 (88.0000)  time: 0.8822  data: 0.8625  max mem: 2928
Test: Total time: 0:00:57 (1.1511 s / it)
* Acc@1 68.782 Acc@5 88.236 loss 1.382
Accuracy of the model on the 50000 test images: 68.8%
Max accuracy: 68.86%
Epoch: [294]  [  0/625]  eta: 3:38:43  lr: 0.000006  min_lr: 0.000006  loss: 2.8759 (2.8759)  class_acc: 0.5391 (0.5391)  weight_decay: 0.0500 (0.0500)  time: 20.9980  data: 19.6595  max mem: 2928
Epoch: [294]  [200/625]  eta: 0:14:27  lr: 0.000005  min_lr: 0.000005  loss: 2.6856 (2.6840)  class_acc: 0.5977 (0.6011)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4792 (2.7672)  time: 1.8199  data: 0.0011  max mem: 2928
Epoch: [294]  [400/625]  eta: 0:07:25  lr: 0.000005  min_lr: 0.000005  loss: 2.6811 (2.6879)  class_acc: 0.5977 (0.5998)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6034 (2.7596)  time: 1.8411  data: 0.0008  max mem: 2928
Epoch: [294]  [600/625]  eta: 0:00:49  lr: 0.000004  min_lr: 0.000004  loss: 2.6740 (2.6836)  class_acc: 0.6016 (0.6002)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4229 (2.7472)  time: 1.8682  data: 0.0011  max mem: 2928
Epoch: [294]  [624/625]  eta: 0:00:01  lr: 0.000004  min_lr: 0.000004  loss: 2.6834 (2.6821)  class_acc: 0.6016 (0.6007)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5511 (2.7357)  time: 0.7413  data: 0.0016  max mem: 2928
Epoch: [294] Total time: 0:19:59 (1.9187 s / it)
Averaged stats: lr: 0.000004  min_lr: 0.000004  loss: 2.6834 (2.6774)  class_acc: 0.6016 (0.6013)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5511 (2.7357)
Test:  [ 0/50]  eta: 0:10:29  loss: 1.3391 (1.3391)  acc1: 66.4000 (66.4000)  acc5: 92.0000 (92.0000)  time: 12.5841  data: 12.5476  max mem: 2928
Test:  [10/50]  eta: 0:01:12  loss: 1.2171 (1.2311)  acc1: 72.8000 (73.0909)  acc5: 91.2000 (90.2546)  time: 1.8086  data: 1.7875  max mem: 2928
Test:  [20/50]  eta: 0:00:38  loss: 1.3790 (1.3766)  acc1: 69.6000 (69.6381)  acc5: 88.8000 (88.9905)  time: 0.7248  data: 0.7048  max mem: 2928
Test:  [30/50]  eta: 0:00:23  loss: 1.5001 (1.3958)  acc1: 68.0000 (69.0839)  acc5: 88.8000 (88.5161)  time: 0.8392  data: 0.8194  max mem: 2928
Test:  [40/50]  eta: 0:00:10  loss: 1.3794 (1.4129)  acc1: 67.2000 (68.3122)  acc5: 88.0000 (88.0781)  time: 0.8178  data: 0.7981  max mem: 2928
Test:  [49/50]  eta: 0:00:00  loss: 1.4099 (1.4162)  acc1: 66.4000 (67.9680)  acc5: 87.2000 (87.8720)  time: 0.4890  data: 0.4699  max mem: 2928
Test: Total time: 0:00:46 (0.9344 s / it)
* Acc@1 68.786 Acc@5 88.232 loss 1.382
Accuracy of the model on the 50000 test images: 68.8%
Max accuracy: 68.86%
Epoch: [295]  [  0/625]  eta: 3:28:53  lr: 0.000004  min_lr: 0.000004  loss: 2.8544 (2.8544)  class_acc: 0.5664 (0.5664)  weight_decay: 0.0500 (0.0500)  time: 20.0535  data: 17.1230  max mem: 2928
Epoch: [295]  [200/625]  eta: 0:14:22  lr: 0.000004  min_lr: 0.000004  loss: 2.6610 (2.6748)  class_acc: 0.5977 (0.6009)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.7520 (2.8132)  time: 1.8582  data: 0.0006  max mem: 2928
Epoch: [295]  [400/625]  eta: 0:07:25  lr: 0.000003  min_lr: 0.000003  loss: 2.7026 (2.6805)  class_acc: 0.5859 (0.5997)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4308 (2.7594)  time: 1.9179  data: 0.0006  max mem: 2928
Epoch: [295]  [600/625]  eta: 0:00:49  lr: 0.000003  min_lr: 0.000003  loss: 2.6588 (2.6791)  class_acc: 0.5977 (0.6001)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4203 (2.7308)  time: 2.0897  data: 0.0007  max mem: 2928
Epoch: [295]  [624/625]  eta: 0:00:01  lr: 0.000003  min_lr: 0.000003  loss: 2.6758 (2.6793)  class_acc: 0.5938 (0.6002)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4045 (2.7226)  time: 0.6885  data: 0.0015  max mem: 2928
Epoch: [295] Total time: 0:20:11 (1.9390 s / it)
Averaged stats: lr: 0.000003  min_lr: 0.000003  loss: 2.6758 (2.6792)  class_acc: 0.5938 (0.6014)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4045 (2.7226)
Test:  [ 0/50]  eta: 0:10:24  loss: 1.3216 (1.3216)  acc1: 67.2000 (67.2000)  acc5: 89.6000 (89.6000)  time: 12.4948  data: 12.4627  max mem: 2928
Test:  [10/50]  eta: 0:01:24  loss: 1.2071 (1.2243)  acc1: 72.8000 (73.1636)  acc5: 90.4000 (90.1091)  time: 2.1116  data: 2.0889  max mem: 2928
Test:  [20/50]  eta: 0:00:50  loss: 1.3845 (1.3711)  acc1: 68.8000 (69.8286)  acc5: 88.8000 (88.9905)  time: 1.1300  data: 1.1098  max mem: 2928
Test:  [30/50]  eta: 0:00:28  loss: 1.4981 (1.3908)  acc1: 67.2000 (69.2129)  acc5: 88.8000 (88.3871)  time: 1.0090  data: 0.9900  max mem: 2928
Test:  [40/50]  eta: 0:00:12  loss: 1.3737 (1.4089)  acc1: 67.2000 (68.4683)  acc5: 87.2000 (87.8829)  time: 0.7443  data: 0.7237  max mem: 2928
Test:  [49/50]  eta: 0:00:01  loss: 1.4152 (1.4127)  acc1: 67.2000 (68.2080)  acc5: 87.2000 (87.7440)  time: 0.9513  data: 0.9306  max mem: 2928
Test: Total time: 0:00:57 (1.1432 s / it)
* Acc@1 68.778 Acc@5 88.162 loss 1.380
Accuracy of the model on the 50000 test images: 68.8%
Max accuracy: 68.86%
Epoch: [296]  [  0/625]  eta: 4:00:00  lr: 0.000003  min_lr: 0.000003  loss: 2.5667 (2.5667)  class_acc: 0.6328 (0.6328)  weight_decay: 0.0500 (0.0500)  time: 23.0405  data: 16.2490  max mem: 2928
Epoch: [296]  [200/625]  eta: 0:14:42  lr: 0.000003  min_lr: 0.000003  loss: 2.6616 (2.6685)  class_acc: 0.5977 (0.6031)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3834 (2.7222)  time: 1.9687  data: 0.0088  max mem: 2928
Epoch: [296]  [400/625]  eta: 0:07:32  lr: 0.000002  min_lr: 0.000002  loss: 2.6974 (2.6692)  class_acc: 0.5977 (0.6028)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4679 (2.6991)  time: 2.0311  data: 0.0284  max mem: 2928
Epoch: [296]  [600/625]  eta: 0:00:49  lr: 0.000002  min_lr: 0.000002  loss: 2.6595 (2.6766)  class_acc: 0.6016 (0.6010)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3063 (2.7455)  time: 2.0483  data: 0.0102  max mem: 2928
Epoch: [296]  [624/625]  eta: 0:00:01  lr: 0.000002  min_lr: 0.000002  loss: 2.6517 (2.6753)  class_acc: 0.6016 (0.6013)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3670 (2.7370)  time: 0.7539  data: 0.0016  max mem: 2928
Epoch: [296] Total time: 0:20:11 (1.9383 s / it)
Averaged stats: lr: 0.000002  min_lr: 0.000002  loss: 2.6517 (2.6778)  class_acc: 0.6016 (0.6014)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3670 (2.7370)
Test:  [ 0/50]  eta: 0:08:50  loss: 1.3327 (1.3327)  acc1: 66.4000 (66.4000)  acc5: 92.0000 (92.0000)  time: 10.6021  data: 10.5676  max mem: 2928
Test:  [10/50]  eta: 0:01:18  loss: 1.2059 (1.2268)  acc1: 72.8000 (73.2364)  acc5: 91.2000 (90.5455)  time: 1.9548  data: 1.9344  max mem: 2928
Test:  [20/50]  eta: 0:00:47  loss: 1.3788 (1.3739)  acc1: 69.6000 (69.7143)  acc5: 88.8000 (89.1429)  time: 1.1154  data: 1.0961  max mem: 2928
Test:  [30/50]  eta: 0:00:25  loss: 1.4987 (1.3939)  acc1: 67.2000 (69.0065)  acc5: 88.8000 (88.7226)  time: 0.9344  data: 0.9151  max mem: 2928
Test:  [40/50]  eta: 0:00:10  loss: 1.3849 (1.4118)  acc1: 67.2000 (68.3512)  acc5: 88.0000 (88.1951)  time: 0.5183  data: 0.4998  max mem: 2928
Test:  [49/50]  eta: 0:00:00  loss: 1.4114 (1.4146)  acc1: 66.4000 (68.0320)  acc5: 87.2000 (88.0480)  time: 0.4720  data: 0.4534  max mem: 2928
Test: Total time: 0:00:45 (0.9118 s / it)
* Acc@1 68.840 Acc@5 88.248 loss 1.380
Accuracy of the model on the 50000 test images: 68.8%
Max accuracy: 68.86%
Epoch: [297]  [  0/625]  eta: 3:10:45  lr: 0.000002  min_lr: 0.000002  loss: 2.7148 (2.7148)  class_acc: 0.5859 (0.5859)  weight_decay: 0.0500 (0.0500)  time: 18.3130  data: 16.6817  max mem: 2928
Epoch: [297]  [200/625]  eta: 0:13:56  lr: 0.000002  min_lr: 0.000002  loss: 2.6638 (2.6785)  class_acc: 0.5977 (0.6011)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4251 (2.6761)  time: 2.1036  data: 0.4350  max mem: 2928
Epoch: [297]  [400/625]  eta: 0:07:22  lr: 0.000002  min_lr: 0.000002  loss: 2.7218 (2.6803)  class_acc: 0.5898 (0.6005)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.3476 (2.7105)  time: 1.9519  data: 0.0009  max mem: 2928
Epoch: [297]  [600/625]  eta: 0:00:49  lr: 0.000002  min_lr: 0.000002  loss: 2.6509 (2.6761)  class_acc: 0.6016 (0.6020)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.9869 (2.7638)  time: 2.1211  data: 0.0010  max mem: 2928
Epoch: [297]  [624/625]  eta: 0:00:01  lr: 0.000002  min_lr: 0.000002  loss: 2.6145 (2.6755)  class_acc: 0.6055 (0.6022)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6785 (2.7563)  time: 0.7132  data: 0.0037  max mem: 2928
Epoch: [297] Total time: 0:20:02 (1.9242 s / it)
Averaged stats: lr: 0.000002  min_lr: 0.000002  loss: 2.6145 (2.6766)  class_acc: 0.6055 (0.6021)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6785 (2.7563)
Test:  [ 0/50]  eta: 0:10:29  loss: 1.3225 (1.3225)  acc1: 67.2000 (67.2000)  acc5: 91.2000 (91.2000)  time: 12.5910  data: 12.5570  max mem: 2928
Test:  [10/50]  eta: 0:01:28  loss: 1.2055 (1.2277)  acc1: 72.8000 (73.0182)  acc5: 91.2000 (90.5455)  time: 2.2065  data: 2.1870  max mem: 2928
Test:  [20/50]  eta: 0:00:53  loss: 1.3894 (1.3742)  acc1: 69.6000 (69.5619)  acc5: 90.4000 (89.1810)  time: 1.2540  data: 1.2355  max mem: 2928
Test:  [30/50]  eta: 0:00:32  loss: 1.4979 (1.3939)  acc1: 68.0000 (68.9290)  acc5: 88.8000 (88.6194)  time: 1.2952  data: 1.2756  max mem: 2928
Test:  [40/50]  eta: 0:00:13  loss: 1.3860 (1.4120)  acc1: 68.0000 (68.2341)  acc5: 88.0000 (88.1561)  time: 0.9179  data: 0.8975  max mem: 2928
Test:  [49/50]  eta: 0:00:01  loss: 1.4148 (1.4150)  acc1: 66.4000 (67.9840)  acc5: 87.2000 (87.9680)  time: 0.8327  data: 0.8112  max mem: 2928
Test: Total time: 0:00:56 (1.1392 s / it)
* Acc@1 68.774 Acc@5 88.224 loss 1.382
Accuracy of the model on the 50000 test images: 68.8%
Max accuracy: 68.86%
Epoch: [298]  [  0/625]  eta: 3:39:39  lr: 0.000002  min_lr: 0.000002  loss: 2.6401 (2.6401)  class_acc: 0.6172 (0.6172)  weight_decay: 0.0500 (0.0500)  time: 21.0868  data: 18.8524  max mem: 2928
Epoch: [298]  [200/625]  eta: 0:13:59  lr: 0.000001  min_lr: 0.000001  loss: 2.6615 (2.6802)  class_acc: 0.6055 (0.6003)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6550 (2.6390)  time: 1.9461  data: 0.4480  max mem: 2928
Epoch: [298]  [400/625]  eta: 0:07:17  lr: 0.000001  min_lr: 0.000001  loss: 2.6597 (2.6818)  class_acc: 0.6055 (0.5996)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5365 (2.7021)  time: 2.0119  data: 0.0005  max mem: 2928
Epoch: [298]  [600/625]  eta: 0:00:48  lr: 0.000001  min_lr: 0.000001  loss: 2.5863 (2.6811)  class_acc: 0.6250 (0.5999)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6566 (2.6847)  time: 2.0652  data: 0.0006  max mem: 2928
Epoch: [298]  [624/625]  eta: 0:00:01  lr: 0.000001  min_lr: 0.000001  loss: 2.6654 (2.6816)  class_acc: 0.5977 (0.5996)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6008 (2.6869)  time: 0.4461  data: 0.0017  max mem: 2928
Epoch: [298] Total time: 0:19:59 (1.9197 s / it)
Averaged stats: lr: 0.000001  min_lr: 0.000001  loss: 2.6654 (2.6778)  class_acc: 0.5977 (0.6015)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6008 (2.6869)
Test:  [ 0/50]  eta: 0:10:08  loss: 1.3397 (1.3397)  acc1: 67.2000 (67.2000)  acc5: 91.2000 (91.2000)  time: 12.1776  data: 12.1515  max mem: 2928
Test:  [10/50]  eta: 0:01:20  loss: 1.2115 (1.2323)  acc1: 72.8000 (73.3818)  acc5: 91.2000 (90.4727)  time: 2.0163  data: 1.9936  max mem: 2928
Test:  [20/50]  eta: 0:00:47  loss: 1.3832 (1.3786)  acc1: 70.4000 (69.9429)  acc5: 89.6000 (89.1429)  time: 1.0541  data: 1.0322  max mem: 2928
Test:  [30/50]  eta: 0:00:27  loss: 1.4972 (1.3980)  acc1: 68.0000 (69.2903)  acc5: 88.8000 (88.6194)  time: 1.0158  data: 0.9955  max mem: 2928
Test:  [40/50]  eta: 0:00:11  loss: 1.3818 (1.4153)  acc1: 68.0000 (68.5659)  acc5: 88.0000 (88.1756)  time: 0.6660  data: 0.6472  max mem: 2928
Test:  [49/50]  eta: 0:00:00  loss: 1.4170 (1.4186)  acc1: 67.2000 (68.2880)  acc5: 87.2000 (87.9520)  time: 0.6053  data: 0.5834  max mem: 2928
Test: Total time: 0:00:47 (0.9452 s / it)
* Acc@1 68.832 Acc@5 88.186 loss 1.385
Accuracy of the model on the 50000 test images: 68.8%
Max accuracy: 68.86%
Epoch: [299]  [  0/625]  eta: 3:18:11  lr: 0.000001  min_lr: 0.000001  loss: 2.3921 (2.3921)  class_acc: 0.6289 (0.6289)  weight_decay: 0.0500 (0.0500)  time: 19.0265  data: 18.4969  max mem: 2928
Epoch: [299]  [200/625]  eta: 0:13:50  lr: 0.000001  min_lr: 0.000001  loss: 2.7380 (2.6751)  class_acc: 0.5898 (0.6017)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5679 (2.6599)  time: 1.7346  data: 0.0006  max mem: 2928
Epoch: [299]  [400/625]  eta: 0:07:18  lr: 0.000001  min_lr: 0.000001  loss: 2.7213 (2.6777)  class_acc: 0.5977 (0.6016)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.6202 (2.7620)  time: 1.8885  data: 0.0006  max mem: 2928
Epoch: [299]  [600/625]  eta: 0:00:48  lr: 0.000001  min_lr: 0.000001  loss: 2.6587 (2.6767)  class_acc: 0.6172 (0.6021)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.4148 (2.7151)  time: 1.9842  data: 0.0005  max mem: 2928
Epoch: [299]  [624/625]  eta: 0:00:01  lr: 0.000001  min_lr: 0.000001  loss: 2.7043 (2.6765)  class_acc: 0.6016 (0.6023)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5488 (2.7135)  time: 0.9958  data: 0.0012  max mem: 2928
Epoch: [299] Total time: 0:19:54 (1.9115 s / it)
Averaged stats: lr: 0.000001  min_lr: 0.000001  loss: 2.7043 (2.6769)  class_acc: 0.6016 (0.6016)  weight_decay: 0.0500 (0.0500)  grad_norm: 2.5488 (2.7135)
Test:  [ 0/50]  eta: 0:10:15  loss: 1.3242 (1.3242)  acc1: 67.2000 (67.2000)  acc5: 92.0000 (92.0000)  time: 12.3172  data: 12.2858  max mem: 2928
Test:  [10/50]  eta: 0:01:20  loss: 1.2083 (1.2299)  acc1: 72.8000 (73.3091)  acc5: 91.2000 (90.4000)  time: 2.0181  data: 1.9960  max mem: 2928
Test:  [20/50]  eta: 0:00:46  loss: 1.3931 (1.3753)  acc1: 68.8000 (69.6000)  acc5: 90.4000 (89.0667)  time: 1.0217  data: 1.0017  max mem: 2928
Test:  [30/50]  eta: 0:00:27  loss: 1.5069 (1.3959)  acc1: 67.2000 (68.8774)  acc5: 88.8000 (88.6452)  time: 1.0302  data: 1.0113  max mem: 2928
Test:  [40/50]  eta: 0:00:12  loss: 1.3897 (1.4142)  acc1: 67.2000 (68.1756)  acc5: 88.0000 (88.1366)  time: 0.9157  data: 0.8972  max mem: 2928
Test:  [49/50]  eta: 0:00:01  loss: 1.4096 (1.4175)  acc1: 67.2000 (67.9520)  acc5: 87.2000 (87.9520)  time: 0.8629  data: 0.8446  max mem: 2928
Test: Total time: 0:00:56 (1.1277 s / it)
* Acc@1 68.808 Acc@5 88.270 loss 1.382
Accuracy of the model on the 50000 test images: 68.8%
Max accuracy: 68.86%
Training time 12:06:45