# Model_zoo **Repository Path**: Arm-China/Model_zoo ## Basic Information - **Project Name**: Model_zoo - **Description**: 安谋科技NPU模型仓库，github镜像仓库： https://github.com/Arm-China/Model_zoo - **Primary Language**: Unknown - **License**: Not specified - **Default Branch**: main - **Homepage**: None - **GVP Project**: No ## Statistics - **Stars**: 3 - **Forks**: 3 - **Created**: 2022-11-04 - **Last Updated**: 2025-10-22 ## Categories & Tags **Categories**: Uncategorized **Tags**: None ## README ## Zhouyi Model Zoo This repo provides a set of AI models for reference used by Zhouyi SDK. | Model | Framework | Input Shape | Model Source | Quant Model | |---------------------------|-----------|---------------------------------------------|---------------------------------------------------------------------------------------------------------------------------------------------------|-------------| | 3d_unet | onnx | [1,1,128,128,128] | https://github.com/mlcommons/inference/tree/master/vision/medical_imaging/3d-unet-kits19 | No | | Baichuan2-7b | onnx | - | https://huggingface.co/baichuan-inc/Baichuan2-7B-Chat | No | | ChatGLM3-6b | hf | [1,256,4096], [1,1,256,512], [1,256] | https://huggingface.co/THUDM/chatglm3-6b | No | | MLP_EDSR | tflite | [1, 300, 500, 3] | https://github.com/mlcommons/mobile_open/tree/main/vision/edsr | No | | MLP_deeplabV3_plus_MV2 | tflite | [1, 512, 512, 3] | https://github.com/mlcommons/mobile_models/blob/main/v0_7/tflite/deeplabv3_mnv2_ade20k_int8.tflite | No | | MLP_mobiledet | tflite | [1, 320, 320, 3] | https://github.com/mlcommons/mobile_open/tree/main/vision/mobiledet | No | | MLP_ssd_mobilenet_v2 | tflite | [1, 300, 300, 3] | https://github.com/mlcommons/mobile_open/tree/main/vision/ssd_mobilenet | No | | Qwen1.5-1.8B | hf | - | https://huggingface.co/Qwen/Qwen1.5-1.8B-Chat | No | | Stable_Diffusion_V1.5 | onnx | [2,4,64,64],[1],[2,77,768] | https://huggingface.co/runwayml/stable-diffusion-v1-5 | No | | ViT_B_16 | pytorch | [1, 3, 224, 224] | https://pytorch.org/vision/stable/models/vision_transformer.html | No | | Whisper_medium_en | onnx | - | https://huggingface.co/openai/whisper-medium.en/tree/main | No | | age_googlenet | onnx | [1, 3, 224, 224] | https://github.com/onnx/models/tree/master/vision/body_analysis/age_gender | No | | age_googlenet | onnx | [1, 3, 224, 224] | https://github.com/onnx/models/tree/master/vision/body_analysis/age_gender | Yes | | alexnet | tflite | [1,28,28,1] | https://github.com/tensorflow/models/tree/archive/research/slim/nets | No | | alexnet | caffe | [1,3,227,227] | https://github.com/BVLC/caffe/tree/master/models/bvlc_alexnet | No | | alexnet | onnx | [1,1,28,28] | https://github.com/tensorflow/models/tree/archive/research/slim/nets | No | | alexnet | tf1 | [1,28,28,1] | https://github.com/tensorflow/models/tree/archive/research/slim/nets | No | | alexnet | pytorch | [2,3,224,224] | https://pytorch.org/vision/stable/models/generated/torchvision.models.alexnet.html#torchvision.models.alexnet | No | | ann | onnx | [1, 3, 1024, 2048] | https://github.com/open-mmlab/mmsegmentation/tree/master/configs/ann | No | | apcnet | onnx | [1, 3, 1024, 2048] | https://github.com/open-mmlab/mmsegmentation/tree/master/configs/apcnet | No | | arcface | onnx | [1, 3, 112, 112] | https://github.com/onnx/models/tree/master/vision/body_analysis/arcface | No | | bisenet_v2 | tf1 | [4, 512, 1024, 3] | https://github.com/MaybeShewill-CV/bisenetv2-tensorflow | No | | caffenet | onnx | [1, 3, 224, 224] | https://github.com/onnx/models/tree/master/vision/classification/caffenet | No | | caffenet | caffe | [10, 3, 227, 227] | https://github.com/BVLC/caffe/tree/master/models/bvlc_reference_caffenet | No | | caffenet | onnx | [1, 3, 224, 224] | https://github.com/onnx/models/tree/master/vision/classification/caffenet | Yes | | centerface | onnx | [1, 3, 640, 640] | https://github.com/ttanzhiqiang/onnx_tensorrt_project | No | | centernet | tf1 | [1, 512, 512, 3] | https://hailo.ai/devzone-model-zoo/object-detection/ | No | | centernet_resnet50 | pytorch | [1, 3, 512, 512] | https://github.com/bubbliiiing/centernet-pytorch | No | | codeformer_256 | onnx | [1,3,256,256] | https://github.com/sczhou/CodeFormer/releases/tag/v0.1.0 | No | | crnn_quant | tflite | [1,64,200,3] | https://ai-benchmark.com/download.html | Yes | | deeplab_v2 | onnx | [1, 3, 513, 513] | https://github.com/kazuto1011/deeplab-pytorch | No | | deeplab_v3 | tflite | [1,257,257,3] | https://github.com/tensorflow/models/tree/archive/research/deeplab | No | | deeplab_v3 | onnx | [1,3,513,513] | https://github.com/tensorflow/models/tree/archive/research/deeplab | No | | deeplab_v3 | tf1 | [1,513,513,3] | https://github.com/tensorflow/models/tree/archive/research/deeplab | No | | deeplab_v3_plus_quant | tflite | [1,1024,1024,3] | https://ai-benchmark.com/download.html | Yes | | deeplab_v3_xception | tf1 | [1,513,513,3] | https://github.com/tensorflow/models/blob/66264b2353aeeca3d4b340908a9590571495d5a6/research/deeplab/g3doc/model_zoo.md | No | | deeplabv3_mobilenetv2 | onnx | [1, 3, 512, 512] | https://github.com/zym1119/DeepLabv3_MobileNetv2_PyTorch | No | | deepspeech_v2 | tf1 | [1,385,161,1] | https://github.com/tensorflow/models/tree/archive/research/deep_speech | No | | deepspeech_v2 | onnx | [1,385,161,1] | https://github.com/tensorflow/models/tree/archive/research/deep_speech | No | | densenet | tflite | [1, 224, 224, 3] | https://www.tensorflow.org/lite/guide/hosted_models?hl=zh-cn | No | | densenet_121 | caffe | [1, 3, 224, 224] | https://github.com/soeaver/caffe-model | No | | densenet_121 | onnx | [1, 3, 224, 224] | https://github.com/onnx/models/tree/master/vision/classification/densenet-121 | No | | densenet_121 | tf2 | [1, 224, 224, 3] | https://www.tensorflow.org/versions/r2.6/api_docs/python/tf/keras/applications | No | | densenet_169 | caffe | [1, 3, 224, 224] | https://github.com/soeaver/caffe-model | No | | densenet_201 | caffe | [1, 3, 224, 224] | https://github.com/soeaver/caffe-model | No | | dilation_8 | caffe | [1, 3, 900, 900] | https://github.com/fyu/dilation | No | | dped_instance_quant | tflite | [1,1024,1536,3] | https://ai-benchmark.com/download.html | Yes | | dped_quant | tflite | [1,1536,2048,3] | https://ai-benchmark.com/download.html | Yes | | dpn_68_extra | caffe | [1, 3, 224, 224] | https://github.com/soeaver/caffe-model | No | | dpn_92 | caffe | [1, 3, 224, 224] | https://github.com/soeaver/caffe-model | No | | drn_38 | onnx | [1, 3, 512, 1024] | https://github.com/fyu/drn | No | | duc | onnx | [1, 3, 800, 800] | https://github.com/onnx/models/tree/master/vision/object_detection_segmentation/duc | No | | efficientdet | tflite | [ 1,512, 512, 3] | https://tfhub.dev/tensorflow/lite-model/efficientdet/lite3/detection/default/1 | No | | efficientnet_b4_quant | tflite | [1,380,380,3] | https://ai-benchmark.com/download.html | Yes | | efficientnet_b5 | tf1 | [1, 456, 456, 3] | https://github.com/tensorflow/tpu/tree/master/models/official/efficientnet | No | | efficientnet_b5 | pytorch | [1, 3, 456, 456] | https://pytorch.org/vision/stable/models/generated/torchvision.models.efficientnet_b5.html#torchvision.models.efficientnet_b5 | No | | efficientnet_b5 | tf2 | [1,456,456,3] | https://www.tensorflow.org/versions/r2.6/api_docs/python/tf/keras/applications | No | | efficientnet_l | tf1 | [1, 300, 300, 3] | https://hailo.ai/devzone-model-zoo/about-object-detection/ | No | | efficientnet_lite | onnx | [1,224,224,3] | https://github.com/onnx/models/tree/master/vision/classification/efficientnet-lite4 | No | | efficientnet_lite | tf1 | [1, 280, 280, 3] | https://hailo.ai/devzone-model-zoo/about-object-detection/ | No | | efficientnet_lite | tflite | [1, 300, 300, 3] | https://tfhub.dev/tensorflow/efficientnet/lite4/classification/2 | No | | enet | caffe | [1, 3, 512, 1024] | https://github.com/TimoSaemann/Enet | No | | erfnet | caffe | [1,3,512,1024] | https://github.com/Yuelong-Yu/ERFNet-Caffe | No | | esrgan_quant | tflite | [1,128,128,3] | https://ai-benchmark.com/download.html | Yes | | face_boxes | onnx | [1,3,1024,1024] | https://github.com/zisianw/FaceBoxes.PyTorch | No | | facenet | tf1 | [1, 160, 160, 3] | https://github.com/davidsandberg/facenet | No | | fast_depth | onnx | [1,3,224,224] | https://github.com/dwofk/fast-depth | No | | faster_rcnn | caffe | [1, 3, 224, 224] | https://github.com/rbgirshick/py-faster-rcnn | No | | fcn | caffe | [1,3,224,224] | https://github.com/shelhamer/fcn.berkeleyvision.org/tree/master/voc-fcn8s-atonce | No | | fcn | pytorch | [1, 3, 224, 224] | https://pytorch.org/vision/stable/models/fcn.html | No | | fcn16s | onnx | [1, 3, 224, 224] | https://github.com/wkentaro/pytorch-fcn/ | No | | fcn32s | onnx | [1, 3, 224, 224] | https://github.com/wkentaro/pytorch-fcn/ | No | | fcn32s | onnx | [1, 3, 224, 224] | https://github.com/wkentaro/pytorch-fcn/ | Yes | | fcn8s | onnx | [1, 3, 224, 224] | https://github.com/wkentaro/pytorch-fcn/ | No | | fsrcnn | onnx | [1, 1, 85, 85] | https://github.com/yjn870/FSRCNN-pytorch | No | | gender_googlenet | onnx | [1, 3, 224, 224] | https://github.com/onnx/models/tree/master/vision/body_analysis/age_gender | No | | googlenet | caffe | [10, 3, 224, 224] | https://github.com/BVLC/caffe/tree/master/models/bvlc_googlenet | No | | googlenet | onnx | [1, 3, 224, 224] | https://github.com/onnx/models/tree/master/vision/classification/inception_and_googlenet/googlenet | No | | googlenet | pytorch | [1, 3, 224, 224] | https://pytorch.org/vision/stable/models/googlenet.html | Yes | | gru_l | tf1 | [1, 49, 10] | https://github.com/UT2UH/ML-KWS-for-ESP32/tree/master/Pretrained_models/GRU | No | | icnet | caffe | [1, 3, 1025, 2049] | https://github.com/hszhao/ICNet | No | | imdn_quant | tflite | [1,1024,1024,3] | https://ai-benchmark.com/download.html | Yes | | inception_resnet_v2 | tflite | [1, 299, 299, 3] | https://www.tensorflow.org/lite/guide/hosted_models?hl=zh-cn | No | | inception_resnet_v2 | caffe | [1, 3,331,331] | https://github.com/soeaver/caffe-model | No | | inception_resnet_v2 | tf1 | [1, 299, 299, 3] | https://github.com/tensorflow/models/tree/master/research/slim#Pretrained | No | | inception_resnet_v2 | tf2 | [1,299,299,3] | https://www.tensorflow.org/api_docs/python/tf/keras/applications/InceptionResNetV2 | No | | inception_v1 | onnx | [1, 3, 224, 224] | https://github.com/onnx/models/tree/master/vision/classification/inception_and_googlenet/inception_v1 | No | | inception_v1 | tf1 | [1, 224, 224, 3] | https://github.com/tensorflow/models/tree/master/research/slim#Pretrained | No | | inception_v1 | tflite | [1, 224, 224, 3] | https://www.tensorflow.org/lite/guide/hosted_models?hl=zh-cn | No | | inception_v1 | onnx | [1,3,224,224] | https://github.com/onnx/models | Yes | | inception_v2 | onnx | [1, 3, 224, 224] | https://github.com/onnx/models/tree/master/vision/classification/inception_and_googlenet/inception_v2 | No | | inception_v2 | tf1 | [1, 224, 224, 3] | https://github.com/tensorflow/models/tree/master/research/slim#Pretrained | No | | inception_v2 | tflite | [1, 224, 224, 3] | https://www.tensorflow.org/lite/guide/hosted_models?hl=zh-cn | No | | inception_v2_ssd | tf1 | [1, 300, 300, 3] | https://github.com/tensorflow/models/blob/master/research/object_detection/g3doc/tf1_detection_zoo.md | No | | inception_v3 | tf2 | [1, 299, 299, 3] | https://www.tensorflow.org/versions/r2.6/api_docs/python/tf/keras/applications | No | | inception_v3 | onnx | [1, 3, 299, 299] | https://github.com/tensorflow/models/tree/archive/research/slim/nets | No | | inception_v3 | tflite | [1, 299, 299, 3] | https://github.com/tensorflow/models/tree/archive/research/slim/nets | No | | inception_v3 | caffe | [1, 3, 299, 299] | https://github.com/soeaver/caffe-model/tree/master/cls | No | | inception_v3 | tf1 | [1, 299, 299, 3] | https://github.com/tensorflow/models/tree/archive/research/slim/nets | No | | inception_v3 | pytorch | [1,3,224,224] | https://pytorch.org/vision/stable/models/generated/torchvision.models.quantization.inception_v3.html#torchvision.models.quantization.inception_v3 | Yes | | inception_v3_quant | tflite | [1, 346, 346, 3] | https://ai-benchmark.com/download.html | Yes | | inception_v4 | tf1 | [1,299,299,3] | https://github.com/tensorflow/models/tree/archive/research/slim/nets | No | | inception_v4 | tflite | [1,299,299,3] | https://github.com/tensorflow/models/tree/archive/research/slim/nets | No | | inception_v4 | onnx | [1,3,299,299] | https://github.com/tensorflow/models/tree/archive/research/slim/nets | No | | inception_v4 | caffe | [1,3,299,299] | https://github.com/soeaver/caffe-model/tree/master/cls | No | | lightface | onnx | [1, 3, 240, 320] | https://hailo.ai/devzone-model-zoo/face-detection/ | No | | llama2-13b | onnx | - | https://huggingface.co/meta-llama/Llama-2-13b-chat | No | | llama2-7b | hf | - | https://huggingface.co/meta-llama/Llama-2-7b-chat | No | | llama2_chn | hf | [1,256,4096], [1,1,256,512], [1,256] | https://huggingface.co/Linly-AI/Chinese-LLaMA-2-7B-hf | No | | lstm_quant | tflite | [1,32,500,1] | https://ai-benchmark.com/download.html | Yes | | maskrcnn | tf1 | [1,1024,1024,3] | https://github.com/matterport/Mask_RCNN | No | | maskrcnn | pytorch | [3,800,800] | https://pytorch.org/vision/main/models/mask_rcnn.html | No | | mixnet | tf1 | [1,224,224,3] | https://github.com/openvinotoolkit/open_model_zoo/tree/master/models/public/mixnet-l | No | | mnasnet | tflite | [1, 224, 224, 3] | https://www.tensorflow.org/lite/guide/hosted_models?hl=zh-cn | No | | mobilebert_quant | tflite | [1,384], [1,384], [1,384] | https://ai-benchmark.com/download.html | Yes | | mobiledet_ssd | tflite | [1, 320, 320, 3] | https://github.com/tensorflow/models/blob/master/research/object_detection/g3doc/tf1_detection_zoo.md | No | | mobiledet_ssd | tflite | [1, 320, 320, 3] | https://github.com/tensorflow/models/blob/master/research/object_detection/g3doc/tf1_detection_zoo.md | Yes | | mobilenet_edgetpu | tflite | [1,224,224,3] | https://github.com/mlcommons/mobile_models/blob/main/v0_7/tflite/mobilenet_edgetpu_224_1.0_uint8.tflite | No | | mobilenet_v1 | caffe | [1, 3, 224, 224] | https://github.com/shicai/MobileNet-Caffe | No | | mobilenet_v1 | tflite | [1, 224, 224, 3] | https://www.tensorflow.org/lite/guide/hosted_models?hl=zh-cn | No | | mobilenet_v1_224 | tf1 | [1, 224, 224, 3] | https://github.com/tensorflow/models/tree/master/research/slim#Pretrained | No | | mobilenet_v1_224 | tflite | [1, 224, 224, 3] | https://github.com/tensorflow/models/blob/master/research/slim/nets/mobilenet_v1.md | No | | mobilenet_v1_ssd | tf1 | [1, 300, 300, 3] | https://github.com/tensorflow/models/blob/master/research/object_detection/g3doc/tf1_detection_zoo.md | No | | mobilenet_v1_ssd | onnx | [1, 3, 416, 416] | https://github.com/onnx/models/tree/master/vision/object_detection_segmentation/ssd-mobilenetv1 | No | | mobilenet_v1_ssd | tflite | [300, 300] | https://tfhub.dev/tensorflow/lite-model/ssd_mobilenet_v1/1/metadata/2 | No | | mobilenet_v2 | caffe | [1,3,224,224] | https://github.com/shicai/MobileNet-Caffe | No | | mobilenet_v2 | tf2 | [1, 224, 224, 3] | https://www.tensorflow.org/versions/r2.6/api_docs/python/tf/keras/applications | No | | mobilenet_v2 | tflite | [1,224,224,3] | https://github.com/tensorflow/models/tree/archive/research/slim/nets/mobilenet | No | | mobilenet_v2 | onnx | [1,3,224,224] | https://github.com/tensorflow/models/tree/archive/research/slim/nets/mobilenet | No | | mobilenet_v2 | tf1 | [1,224,224,3] | https://github.com/tensorflow/models/tree/archive/research/slim/nets/mobilenet | No | | mobilenet_v2 | onnx | [1,3,224,224] | https://github.com/tensorflow/models/tree/archive/research/slim/nets/mobilenet | Yes | | mobilenet_v2_b8_quant | tflite | [8,224,224,3] | https://ai-benchmark.com/download.html | Yes | | mobilenet_v2_quant | tflite | [1,224,224,3] | https://ai-benchmark.com/download.html | Yes | | mobilenet_v2_ssd | onnx | [1,300,300,3] | https://github.com/tensorflow/models/tree/archive/research/object_detection/models | No | | mobilenet_v2_ssd | caffe | [1,3,300,300] | https://github.com/chuanqi305/MobileNet-SSD | No | | mobilenet_v2_ssd | tf1 | [1,300,300,3] | https://github.com/tensorflow/models/tree/archive/research/object_detection/models | No | | mobilenet_v2_ssd_lite | tf1 | [1, 300, 300, 3] | https://github.com/tensorflow/models/blob/master/research/object_detection/g3doc/tf1_detection_zoo.md | No | | mobilenet_v3 | tf2 | [1, 224, 224, 3] | https://www.tensorflow.org/versions/r2.6/api_docs/python/tf/keras/applications | No | | mobilenet_v3 | tf1 | [1, 224, 224, 3] | https://github.com/tensorflow/models/tree/master/research/slim/nets/mobilenet | No | | mobilenet_v3 | pytorch | [1, 3, 224, 224] | https://pytorch.org/vision/stable/models/mobilenetv3.html | Yes | | mobilenet_v3 | tflite | [1,224,224,3] | https://github.com/tensorflow/models/tree/master/research/slim/nets/mobilenet | No | | mobilenet_v3_b4_quant | tflite | [4,512,512,3] | https://ai-benchmark.com/download.html | Yes | | mobilenet_v3_quant | tflite | [1,512,512,3] | https://ai-benchmark.com/download.html | Yes | | mosaic | tflite | [1, 1024, 2048, 3] | https://tfhub.dev/google/lite-model/mosaic/mobilenetmultiavgseg/1 | No | | mtcnn_o | caffe | [1,3,48,48] | https://github.com/CongWeilin/mtcnn-caffe | No | | mtcnn_p | caffe | [1,3,12,12] | https://github.com/CongWeilin/mtcnn-caffe | No | | mtcnn_r | caffe | [1,3,24,24] | https://github.com/CongWeilin/mtcnn-caffe | No | | mv3_depth_quant | tflite | [1,1024,1536,3] | https://ai-benchmark.com/download.html | Yes | | nasnet_large | tf1 | [1, 331, 331,3] | https://www.tensorflow.org/lite/guide/hosted_models?hl=zh-cn | No | | nasnet_mobile | tf1 | [1, 224, 224, 3] | https://www.tensorflow.org/lite/guide/hosted_models?hl=zh-cn | No | | nasnet_mobile | tflite | [1, 224, 224, 3] | https://www.tensorflow.org/lite/guide/hosted_models?hl=zh-cn | No | | nasnet_mobile | tf2 | [1, 224, 224, 3] | https://www.tensorflow.org/versions/r2.6/api_docs/python/tf/keras/applications | No | | peleenet | caffe | [1,3,224,224] | https://github.com/Robert-JunWang/PeleeNet/tree/master/caffe | No | | petr | pytorch | [6, 3, 320, 800] | https://github.com/megvii-research/PETR | No | | pnasnet_large | tf1 | [1, 331, 331, 3] | https://github.com/tensorflow/models/tree/master/research/slim#Pretrained | No | | pointnet | onnx | [32, 9, 4096] | https://github.com/yanx27/Pointnet_Pointnet2_pytorch | No | | poly_lanenet | onnx | [1, 3, 640, 640] | https://hailo.ai/devzone-model-zoo/driving-lane-detection/ | No | | pspnet | onnx | [1, 3, 1024, 2048] | https://github.com/open-mmlab/mmsegmentation/tree/master/configs/pspnet | No | | punet_quant | tflite | [1,544,960,4] | https://ai-benchmark.com/download.html | Yes | | pynet_quant | tflite | [1,512,512,3] | https://ai-benchmark.com/download.html | Yes | | regnet_x | onnx | [1, 3, 224, 224] | https://hailo.ai/devzone-model-zoo/about-object-detection/ | No | | resnet_18 | onnx | [1, 3, 224, 224] | https://github.com/onnx/models/tree/main/vision/classification/resnet/model | No | | resnet_34_ssd | onnx | [1, 3, 1200, 1200] | https://github.com/mlcommons/inference/tree/master/vision/classification_and_detection | No | | resnet_34_ssd | tf1 | [1, 1200, 1200, 3] | https://github.com/mlcommons/inference/tree/r0.5/v0.5/classification_and_detection | No | | resnet_50_v1_5 | tf1 | [1, 224, 224, 3] | https://zenodo.org/ | No | | resnet_v1_101 | tf2 | [1, 224, 224, 3] | https://www.tensorflow.org/versions/r2.6/api_docs/python/tf/keras/applications | No | | resnet_v1_101 | tflite | [1,224,224,3] | https://github.com/tensorflow/models/tree/archive/research/slim/nets | No | | resnet_v1_101 | caffe | [1,3,224,224] | https://github.com/SnailTyan/caffe-model-zoo | No | | resnet_v1_101 | onnx | [1,3,224,224] | https://github.com/tensorflow/models/tree/archive/research/slim/nets | No | | resnet_v1_101 | tf1 | [1,224,224,3] | https://github.com/tensorflow/models/tree/archive/research/slim/nets | No | | resnet_v1_152 | tf1 | [1, 224, 224, 3] | https://github.com/tensorflow/models/tree/master/research/slim#Pretrained | No | | resnet_v1_50 | onnx | [1,3,224,224] | https://github.com/tensorflow/models/tree/archive/research/slim/nets | No | | resnet_v1_50 | tf2 | [1, 224, 224, 3] | https://www.tensorflow.org/versions/r2.6/api_docs/python/tf/keras/applications | No | | resnet_v1_50 | tflite | [1,224,224,3] | https://github.com/tensorflow/models/tree/archive/research/slim/nets | No | | resnet_v1_50 | caffe | [1,3,224,224] | https://github.com/SnailTyan/caffe-model-zoo | No | | resnet_v1_50 | pytorch | [1, 3, 224, 224] | https://pytorch.org/vision/stable/models/resnet.html | No | | resnet_v1_50 | tf1 | [1,224,224,3] | https://github.com/tensorflow/models/tree/archive/research/slim/nets | No | | resnet_v1_50 | onnx | [1,3,224,224] | https://github.com/tensorflow/models/tree/archive/research/slim/nets | Yes | | resnet_v1_50 | pytorch | [1, 3, 224, 224] | https://pytorch.org/vision/stable/models/resnet.html | Yes | | resnet_v1_ssd | tf1 | [1, 640, 640, 3] | https://github.com/tensorflow/models/blob/master/research/object_detection/g3doc/tf1_detection_zoo.md | No | | resnet_v2_101 | tflite | [1, 299, 299, 3] | https://www.tensorflow.org/lite/guide/hosted_models?hl=zh-cn | No | | resnet_v2_101 | caffe | [1, 3, 448, 448] | https://github.com/soeaver/caffe-model | No | | resnet_v2_101 | tf1 | [1, 299, 299, 3] | https://github.com/tensorflow/models/tree/master/research/slim#Pretrained | No | | resnet_v2_101 | tf2 | [1, 224, 224, 3] | https://www.tensorflow.org/versions/r2.6/api_docs/python/tf/keras/applications | No | | resnet_v2_152 | caffe | [1, 3, 224, 224] | https://github.com/soeaver/caffe-model | No | | resnet_v2_152 | tf1 | [1, 299, 299, 3] | https://github.com/tensorflow/models/tree/master/research/slim#Pretrained | No | | resnet_v2_50 | tf2 | [1, 224, 224, 3] | https://www.tensorflow.org/versions/r2.6/api_docs/python/tf/keras/applications | No | | resnet_v2_50 | tf1 | [1, 299, 299, 3] | https://github.com/tensorflow/models/tree/master/research/slim#Pretrained | No | | resnext_101 | onnx | [1, 3, 224, 224] | https://github.com/Cadene/pretrained-models.pytorch | No | | resnext_101 | caffe | [1, 3, 224, 224] | https://www.deepdetect.com/models/resnext/ | No | | resnext_50 | caffe | [1, 3, 224, 224] | https://github.com/soeaver/caffe-model | No | | retinanet | onnx | [1, 3, 480, 640] | https://github.com/onnx/models/tree/master/vision/object_detection_segmentation/retinanet | No | | rnn_t_decoder | onnx | [1, 1],[1, 2, 320],[1, 2, 320],[1, 1, 1024] | https://github.com/mlcommons/inference/tree/master/speech_recognition/rnnt | No | | rnn_t_encoder | onnx | [1,249,240] | https://github.com/mlcommons/inference/tree/master/speech_recognition/rnnt | No | | se_inception | caffe | [1, 3, 224, 224] | https://github.com/openvinotoolkit/open_model_zoo/blob/master/models/public/se-inception/README.md | No | | se_resnet_101 | caffe | [1, 3, 224, 224] | https://github.com/hujie-frank/SENet | No | | se_resnet_50 | tf1 | [1, 224, 224, 3] | https://github.com/HiKapok/TF-SENet | No | | se_resnet_50 | caffe | [1, 3, 225, 225] | https://github.com/soeaver/caffe-model | No | | se_resnext_50 | tf1 | [1, 224, 224, 3] | https://github.com/HiKapok/TF-SENet | No | | sea_former | onnx | [1, 3, 512, 1024] | https://github.com/fudan-zvg/SeaFormer/tree/main/seaformer-seg | No | | shufflenet_v1 | onnx | [1, 3, 224, 224] | https://github.com/onnx/models/tree/master/vision/classification/shufflenet | No | | shufflenet_v2 | onnx | [1,3,224,224] | https://github.com/TropComplique/shufflenet-v2-tensorflow | No | | shufflenet_v2 | caffe | [1,3,224,224] | https://github.com/Ewenwan/ShuffleNet-2 | No | | shufflenet_v2 | pytorch | [1,3,224,224] | https://pytorch.org/vision/stable/models/generated/torchvision.models.shufflenet_v2_x1_0.html#torchvision.models.shufflenet_v2_x1_0 | No | | shufflenet_v2 | tf1 | [1,224,224,3] | https://github.com/TropComplique/shufflenet-v2-tensorflow | No | | shufflenet_v2 | onnx | [1,3,224,224] | https://github.com/TropComplique/shufflenet-v2-tensorflow | Yes | | shufflenet_v2 | pytorch | [1, 3, 224, 224] | https://pytorch.org/vision/stable/models/shufflenetv2.html | Yes | | shufflenet_v2 | tflite | [1,224,224,3] | https://github.com/TropComplique/shufflenet-v2-tensorflow | No | | sne_roadseg | onnx | [1,3,384,1248] | https://github.com/hlwang1124/SNE-RoadSeg | No | | sphereface | caffe | [1, 3, 112, 96] | https://github.com/openvinotoolkit/open_model_zoo/blob/master/models/public/Sphereface/README.md | No | | squeezenet | caffe | [10, 3, 227, 227] | https://github.com/forresti/SqueezeNet | No | | squeezenet | onnx | [1, 3, 224, 224] | https://github.com/onnx/models/tree/master/vision/classification/squeezenet | No | | squeezenet | tflite | [1, 224, 224, 3] | https://www.tensorflow.org/lite/guide/hosted_models?hl=zh-cn | No | | squeezenet | onnx | [1, 3, 224, 224] | https://github.com/onnx/models/tree/master/vision/classification/squeezenet | Yes | | srcnn | tf1 | [1, 33, 33, 1] | https://github.com/tegg89/SRCNN-Tensorflow | No | | srgan_quant | tflite | [1,256,256,3] | https://ai-benchmark.com/download.html | Yes | | ssd | onnx | [1, 3, 1200, 1200] | https://github.com/onnx/models/tree/master/vision/object_detection_segmentation/ssd | No | | stacked_hourglass | tf1 | [1, 256, 256, 3] | https://github.com/yuanyuanli85/Stacked_Hourglass_Network_Keras | No | | super_resolution | onnx | [1, 1, 224, 224] | https://github.com/onnx/models/tree/master/vision/super_resolution/sub_pixel_cnn_2016 | No | | swin_transformer | onnx | [1,3,384,384] | https://github.com/microsoft/Swin-Transformer | No | | swin_transformer_base | onnx | [1,3,224,224] | https://github.com/microsoft/Swin-Transformer/blob/main/MODELHUB.md | No | | swin_transformer_tiny_224 | pytorch | [1, 3, 224, 224] | https://pytorch.org/vision/stable/models/swin_transformer.html | No | | transformer_official | tf1 | [1, 32] | https://github.com/Kyubyong/transformer | No | | unet_bio | tf1 | [1,256, 256,1] | https://github.com/zhixuhao/unet | No | | unet_quant | tflite | [1,1024,1024,3] | https://ai-benchmark.com/download.html | Yes | | vgg_16 | caffe | [1,3,224,224] | https://gist.github.com/ksimonyan/211839e770f7b538e2d8#file-readme-md | No | | vgg_16 | onnx | [1,3,224,224] | https://github.com/tensorflow/models/tree/archive/research/slim/nets | No | | vgg_16 | tflite | [1,224,224,3] | https://github.com/tensorflow/models/tree/archive/research/slim/nets | No | | vgg_16 | tf2 | [1, 224, 224, 3] | https://www.tensorflow.org/versions/r2.6/api_docs/python/tf/keras/applications | No | | vgg_16 | pytorch | [1, 3, 224, 224] | https://pytorch.org/vision/stable/models/vgg.html | No | | vgg_16 | tf1 | [1,224,224,3] | https://github.com/tensorflow/models/tree/archive/research/slim/nets | No | | vgg_16 | onnx | [1,3,224,224] | https://github.com/tensorflow/models/tree/archive/research/slim/nets | Yes | | vgg_19 | tf1 | [1, 224, 224, 3] | https://github.com/tensorflow/models/tree/master/research/slim#Pretrained | No | | vgg_19 | tf2 | [1, 224, 224, 3] | https://www.tensorflow.org/versions/r2.6/api_docs/python/tf/keras/applications | No | | vgg_cnn_s | caffe | [10, 3, 224, 224] | https://gist.github.com/ksimonyan/fd8800eeb36e276cd6f9 | No | | vgg_quant | tflite | [1,256,256,1] | https://ai-benchmark.com/download.html | Yes | | vgg_ssd | caffe | [1, 3, 300, 300] | https://github.com/weiliu89/caffe/tree/ssd | No | | vision_transformer | onnx | [1,3,224,224] | https://github.com/jeonsworld/ViT-pytorch | No | | vsr_quant | tflite | [1,540,960,3] | https://ai-benchmark.com/download.html | Yes | | wavenet | tf1 | [1,390,23] | https://github.com/buriburisuri/speech-to-text-wavenet | No | | wavenet | onnx | [1,390,23] | https://github.com/buriburisuri/speech-to-text-wavenet | No | | xception | caffe | [1, 3, 299, 299] | https://github.com/soeaver/caffe-model | No | | xception | tf2 | [1, 299, 299, 3] | https://www.tensorflow.org/versions/r2.6/api_docs/python/tf/keras/applications | No | | xlsr_quant | tflite | [1,360,640,3] | https://ai-benchmark.com/download.html | Yes | | yolact_regnetx | onnx | [1, 3, 512, 512] | https://hailo.ai/devzone-model-zoo/instance-segmentation/ | No | | yolo_v1_tiny | caffe | [1, 3, 448, 448] | https://github.com/xingwangsfu/caffe-yolo | No | | yolo_v2 | onnx | [1, 3, 416, 416] | https://github.com/onnx/models/tree/master/vision/object_detection_segmentation/yolov2-coco | No | | yolo_v2_416 | caffe | [1,3,416,416] | https://github.com/tsingjinyun/caffe-yolov2 | No | | yolo_v2_416 | onnx | [1,3,416,416] | https://github.com/wojciechmo/yolo2 | No | | yolo_v2_416 | tf1 | [1,416,416,3] | https://github.com/wojciechmo/yolo2 | No | | yolo_v2_tiny | onnx | [1, 3, 416, 416] | https://github.com/onnx/models/tree/master/vision/object_detection_segmentation/tiny-yolov2 | No | | yolo_v3 | tf1 | [1, 416, 416, 3] | https://github.com/qqwweee/keras-yolo3 | No | | yolo_v3 | caffe | [1,3,608,608] | https://github.com/foss-for-synopsys-dwc-arc-processors/synopsys-caffe-models/tree/master/caffe_models/yolo_v3 | No | | yolo_v3_tiny | onnx | [1,416,416,3] | https://github.com/onnx/models/tree/main/vision/object_detection_segmentation/tiny-yolov3 | No | | yolo_v4 | onnx | [1, 416, 416, 3] | https://github.com/onnx/models/tree/master/vision/object_detection_segmentation/yolov4 | No | | yolo_v4 | tf1 | [1,416,416,3] | https://github.com/hunglc007/tensorflow-yolov4-tflite | No | | yolo_v4 | pytorch | [1, 3, 416, 416] | https://github.com/bubbliiiing/yolov4-pytorch/tree/master | No | | yolo_v4_tiny | pytorch | [1, 3, 416, 416] | https://github.com/bubbliiiing/yolov4-tiny-pytorch | No | | yolo_v4_tiny | tflite | [1,416,416,3] | https://github.com/hunglc007/tensorflow-yolov4-tflite/tree/master | No | | yolo_v4_tiny_quant | tflite | [1,416,416,3] | https://ai-benchmark.com/download.html | Yes | | yolo_v5 | onnx | [1, 3, 640, 640] | https://github.com/ultralytics/yolov5 | No | | yolo_v5 | tflite | [1, 640, 640, 3] | https://github.com/ultralytics/yolov5 | No | | yolo_v5_s | onnx | [1, 3, 640, 640] | https://github.com/ultralytics/yolov5/ | No | | yolo_v5m | onnx | [1,3,640,640] | https://github.com/DefTruth/lite.ai.toolkit | No | | yolo_v5s | pytorch | [1,3,640,640] | https://github.com/ultralytics/yolov5/ | No | | yolo_v5s6 | onnx | [1, 3, 640, 640] | https://github.com/ultralytics/yolov5/releases/download/v7.0/yolov5s6.pt | No | | yolo_v6s | onnx | [1,3,640,640] | https://github.com/DefTruth/lite.ai.toolkit | No | | yolo_v8_seg | onnx | [1, 3, 640, 640] | https://docs.ultralytics.com/tasks/segment/#models | No | | yolox_l | onnx | [1,3,640,640] | https://github.com/Megvii-BaseDetection/YOLOX | No | | yolox_s | onnx | [1, 3, 640, 640] | https://github.com/Megvii-BaseDetection/YOLOX/tree/main/demo/ONNXRuntime | No | | zfnet_512 | onnx | [1, 3, 224, 224] | https://github.com/onnx/models/tree/master/vision/classification/zfnet-512 | No |   #### **FTP model download (Suggest FTP tool [FileZilla](https://filezilla-project.org/))** - `Host`: sftp://sftp01.armchina.com - `Account`: zhouyi.armchina - `Password`: 114r3cJd   ## About Zhouyi NPU Zhouyi NPU IP is a high efficiency and high performance AI processor that supports wide range of AI application, including vision, audio/speech and NLP. Zhouyi NPU offers fixed function accelerator that speeds up general operations like convolution and activation, meanwhile Zhouyi NPU offers flexible programmability to support new operations or customized operations. Zhouyi NPU IP has been deployed in a number of market segments, including surveillance, automotive, IOT, etc. With proven peformance and scalablility, Zhouyi NPU IP covers 0.25TOPS to hunderds of TOPS, target market from edge computing to data center.   ## About Zhouyi SDK Zhouyi SDK is a full stack of platform that provides fast time-to-market development and deployment capability for user. ![image](https://user-images.githubusercontent.com/85479712/198521602-49e13a31-bb49-424f-b782-5108274d63c3.png) ### Highlighted feature: - Support various framework > Supports TensorFlow, TensorFlow Lite, Caffe, Pytorch and ONNX, extend to support orther framework by IR transition - Graph optimization > Many graph optimization techniques are used to reduce network complexity, like conv+batchnorm+relu fusion, to make graph run fast and efficiently on Zhouyi hardware. - Quantization > Support training-aware quantization model and post-training quantization method. Natively provides advanced-quantization techniques, also provide interface for user customized quantization parameter. - IR (intermediate representation) > Public Zhouyi IR, and can be a standard for other framework to support - Customization > Provide plugin interface for user easy to develop custom operations and integrate into SDK - Simulator > Provide instruction-level simulator, easy for user to evaluate model before porting to hardware