RepVGG

This implements training of RepVGG on the imagenet dataset, mainly modified from DingXiaoH/RepVGG.

RepVGG Detail

For details, see (https://github.com/DingXiaoH/RepVGG)

Requirements

  • Download the ImageNet dataset refet(https://github.com/DingXiaoH/RepVGG)
  • pip3.7 install -r requirements.txt
    • Note: pillow recommends installing a newer version. If the corresponding torchvision version cannot be installed directly, you can use the source code to install the corresponding version. The source code reference link: https://github.com/pytorch/vision Suggestion: the pillow is 9.1.0 and the torchvision is 0.6.0

Training

To train a model, run train.py with the desired model architecture and the path to the imagenet dataset:

# training 1p performance
bash test/train_performance_1p.sh --data_path=real_data_path

# training 8p accuracy
bash test/train_full_8p.sh --data_path=real_data_path

# training 8p performance
bash test/train_performance_8p.sh --data_path=real_data_path

#test 8p accuracy
bash test/train_eval_8p.sh --data_path=real_data_path

# finetune
bash test/train_finetune_1p.sh --data_path=real_data_path

# Online inference demo
python demo.py
 

The output log of the above script will be saved in the current folder.And the specific output log name of each script refers to the script content

RepVGG training result

batch size 256:

名称 iou fps
GPU-1p - -
GPU-8p - 3700
NPU-1p - -
NPU-8p 72.08 600

batch size 2048:

名称 iou fps
GPU-1p - -
GPU-8p 69.60 6578
NPU-1p - -
NPU-8p 69.67 3265

batch size 4096:

名称 iou fps
GPU-1p - -
GPU-8p 69.41 8010
NPU-1p - -
NPU-8p 69.27 8596