README.md

Train base model

How to Run

1. prepare imagenet100 dataset

   pip install datasets matplotlib opencv-python
   mkdir dataset
   wget https://huggingface.co/datasets/ilee0022/ImageNet100/resolve/main/label2text.json -P dataset
   

2. train resnet18 with imagenet100 dataset

   python train.py
   // 'best_model.pth' will be generated in checkpoint directory.
   

3. check a trained resnet18 with imagenet100 dataset

   python infer.py
   

• fp16
  Gpu Mem: 286M
  [TRT_E] Test Top-1 Accuracy: 84.58%
  [TRT_E] Test Top-5 Accuracy: 97.20%
  [TRT_E] 10000 iterations time: 24.6143 [sec]
  [TRT_E] Average FPS: 406.27 [fps]
  [TRT_E] Average inference time: 2.46 [msec]