Open NdaAzr opened 4 years ago
I am wondering why you used NetC.zero_grad() in train.py, however usually zero_grad() function should set on the optimizer? Could you please explain this. It would be great if there were some more comments in the code. Thank you
I am wondering why you used NetC.zero_grad() in train.py, however usually zero_grad() function should set on the optimizer? Could you please explain this. It would be great if there were some more comments in the code. Thank you