how to add middle layers' activation loss functions? #340
Comments
Hi @brisker , Sorry for the late response. Let us know if you have more questions. Cheers, |
@levzlotnik |
@levzlotnik |
Hi @brisker, (Note that I edited your last comment to remove irrelevant people you tagged) To go back to your original question, there could be some confusion I'd like to clear up. If you notice, the yaml file is named "alexnet_bn". That is to say, the intention is to run this not on the original AlexNet, but on a modified AlexNet with batch norm layers. We have it implemented here. All of this wasn't detailed in the yaml files - that's my fault. I pushed updates to both the base FP32 yaml and the DoReFa yaml with details on how to run it and the results I got. Please check those out. Regarding your question on MaxPool - In general the answer is yes, you should replace MaxPool with something that does quant --> maxpool --> quant. Then you define a function that will create this new module and return it, and then add that function to the "replacement factory". Similar to what we do with ReLU in |
Closing due to inactivity. Please reopen if needed. |
@nzmora |
Hi @brisker , You could define "regularization" policies for these layers that will calculate the Cheers, |
@levzlotnik |
Hi @brisker ,
Where |
@levzlotnik |
1 similar comment
@levzlotnik |
If I want to train a dorefa-Net quantization on alexnet , is the command line like this ?
python compress_classifier.py -a alexnet /ImageNet_Share/ --compress=/distiller/examples/quantization/quant_aware_train/alexnet_bn_dorefa.yaml
We do not need to modify the code in
compress_classifier.py
to do this training?The text was updated successfully, but these errors were encountered: