Skip to content

Latest commit

 

History

History
34 lines (24 loc) · 637 Bytes

README.md

File metadata and controls

34 lines (24 loc) · 637 Bytes

tmo_ptq

How to Run

  1. prepare imagenet100 dataset
cd ..
mkdir datasets

// download imagenet100 dataset from kaggle (see below)
  1. train resnet18 with imagenet100 dataset
cd base_model
python train.py
// 'best_model.pth' will be generated in checkpoint directory.
  1. Post-training quantization (PTQ)
cd tmo_ptq
python ptq_onnx_export.py
python onnx2trt.py
// a file 'resnet18_int8_ptq_bf.engine' will be generated in engine directory.

Reference