Optimization API

optimizer/convert_compare.py: To convert a pytorch model to a tflite model and compare the difference between their outputs

optimizer/match.py:

  • Match a NAS model to a original model

    • fix_nasyolo, fix_nasbackbone

  • Match two pytorch models with same structure, but different hyperparameters (ex: input channels, output channels)

    • match_nas, match

  • Extract Yolov5’s backbone from an original model

    • extract_backbone

optimizer/model_evaluation.py: Evaluate the inference time, network parameters and flops of a specific model

optimizer/loss.py: Calculating the loss between teacher model and student model

optimizer/prune.py: Call the function prune to prune models