/v4tiny_pt_quant

quantization for yolo with xilinx/vitis-ai-pytorch

Primary LanguagePython

v4tiny_pt_quant

Quantization for yolo with xilinx/vitis-ai-pytorch

Tool

Model

  • Model: yolov4-tiny pruned with this project or other yolo models
  • Dataset: coco2017, with darknet format
  • Input format: .cfg + .pt/.weights

Errors

  • Error unfixed: aten_op 'view' parse failed(the layout of activation of reshape is ambiguous) when run with python quant.py --quant_mode calib