910A qwen2-7B 推理加载权重报错

尝试使用910A推理 qwen2-7B，qwen1.5-7B推理已跑通
mindformers                   1.0.2
mindpet                       1.0.3
mindspore                     2.2.14
Python 3.9.18
CANN 7.0.0
驱动23.0.0
使用分支1.0.a
双卡推理报错RuntimeError: For 'load_param_into_net', model.layers.0.attention.wk.weight in the argument 'net' should have the same shape as model.layers.0.attention.wk.weight in the argument 'parameter_dict'. But got its shape (1792, 3584) in the argument 'net' and shape (256, 3584) in the argument 'parameter_dict'.May you need to check whether the checkpoint you loaded is correct or the batch size and so on in the 'net' and 'parameter_dict' are same.
使用配置文件
seed: 0
output_dir: './output' # path to save checkpoint/strategy
load_checkpoint: '/home/mwd/data/Qwen2-7B'
src_strategy_path_or_dir: ''
auto_trans_ckpt: True  # If true, auto transform load_checkpoint to load in distributed model
only_save_strategy: False
resume_training: False
use_parallel: True
run_mode: 'predict'

# trainer config
trainer:
  type: CausalLanguageModelingTrainer
  model_name: 'qwen1_5_7b'

# dataset
train_dataset: &train_dataset
  data_loader:
    type: MindDataset
    dataset_dir: ""
    shuffle: True
  input_columns: ["input_ids", "labels", "attention_mask"]
  num_parallel_workers: 8
  python_multiprocessing: False
  drop_remainder: True
  batch_size: 1
  repeat: 1
  numa_enable: False
  prefetch_size: 1
train_dataset_task:
  type: CausalLanguageModelDataset
  dataset_config: *train_dataset

# runner config
runner_config:
  epochs: 5
  batch_size: 1
  sink_mode: True
  sink_size: 2
runner_wrapper:
  type: MFTrainOneStepCell
  scale_sense:
    type: DynamicLossScaleUpdateCell
    loss_scale_value: 65536
    scale_factor: 2
    scale_window: 1000
  use_clip_grad: True

# optimizer
optimizer:
  type: FP32StateAdamWeightDecay
  beta1: 0.9
  beta2: 0.95
  eps: 1.e-6
  weight_decay: 0.1

# lr sechdule
lr_schedule:
  type: CosineWithWarmUpLR
  learning_rate: 1.e-5
  warmup_ratio: 0.01
  total_steps: -1 # -1 means it will load the total steps of the dataset

# callbacks
callbacks:
  - type: MFLossMonitor
  - type: CheckpointMointor
    prefix: "qwen2"
    save_checkpoint_steps: 10000
    keep_checkpoint_max: 3
    integrated_save: False
    async_save: False
  - type: ObsMonitor

# default parallel of device num = 8 for Atlas 800T A2
parallel_config:
  data_parallel: 1
  model_parallel: 2
  pipeline_stage: 1
  micro_batch_num: 1
  vocab_emb_dp: False
  gradient_aggregation_group: 4
# when model parallel is greater than 1, we can set micro_batch_interleave_num=2, that may accelerate the train process.
micro_batch_interleave_num: 1

# recompute config
recompute_config:
  recompute: False
  select_recompute: False
  parallel_optimizer_comm_recompute: False
  mp_comm_recompute: False
  recompute_slice_activation: False

model:
  model_config:
    type: LlamaConfig
    batch_size: 1
    seq_length: 2048
    hidden_size: 3584
    num_layers: 28
    num_heads: 28
    vocab_size: 152064
    intermediate_size: 18944
    qkv_has_bias: True
    rms_norm_eps: 1.0e-6
    theta: 1000000.0
    emb_dropout_prob: 0.0
    eos_token_id: 151643
    pad_token_id: 151643
    bos_token_id: 151643
    compute_dtype: "float16"
    layernorm_compute_type: "float32"
    softmax_compute_type: "float16"
    rotary_dtype: "float16"
    param_init_type: "float16"
    use_past: True
    use_flash_attention: False
    use_paged_attention: False
    # block_size: 16
    # num_blocks: 512
    use_past_shard: False
    offset: 0
    checkpoint_name_or_path: ""
    repetition_penalty: 1
    max_decode_length: 512
    top_k: 0
    top_p: 0.8
    do_sample: False

# configuration items copied from Qwen
    rotary_pct: 1.0
    rotary_emb_base: 10000
    kv_channels: 128

arch:
    type: LlamaForCausalLM

processor:
  return_tensors: ms
  tokenizer:
    model_max_length: 32768
    vocab_file: "../../data/Qwen2-7B/vocab.json"
    merges_file: "../../data/Qwen2-7B/merges.txt"
    unk_token: "<|endoftext|>"
    eos_token: "<|endoftext|>"
    pad_token: "<|endoftext|>"
    type: Qwen2Tokenizer
  type: Qwen2Processor

# mindspore context init config
context:
  mode: 0 #0--Graph Mode; 1--Pynative Mode
  device_target: "Ascend"
  enable_graph_kernel: False
  graph_kernel_flags: "--disable_expand_ops=Softmax,Dropout --enable_parallel_fusion=true --reduce_fuse_depth=8 --enable_auto_tensor_inplace=true"
  ascend_config:
    precision_mode: "must_keep_origin_dtype"
  max_call_depth: 10000
  max_device_memory: "31GB"
  save_graphs: False
  save_graphs_path: "./graph"
  device_id: 0

# parallel context config
parallel:
  parallel_mode: 1 # 0-data parallel, 1-semi-auto parallel, 2-auto parallel, 3-hybrid parallel
  gradients_mean: False
  enable_alltoall: False
  full_batch: True
  search_mode: "sharding_propagation"
  enable_parallel_optimizer: True
  strategy_ckpt_config:
    save_file: "./ckpt_strategy.ckpt"
    only_trainable_params: False
  parallel_optimizer_config:
    gradient_accumulation_shard: False
    parallel_optimizer_threshold: 64

GVP MindSpore/mindformers

内容风险标识

评论 (3)

GVPMindSpore/mindformers

内容风险标识