cybertronai/autograd-hacks

Issues

• ResNet support

  #2 opened 5 years ago by oljike
  3

• Autograd_Hacks for multiple GPU usage

  #15 opened 2 years ago by amirozhan
  1

• UserWarning: Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.

  #10 opened 4 years ago by slerman12
  3

• How to run calculate the gradient per sample in a loop?

  #13 opened 3 years ago by cijerezg
  0

• Extend to support Conv1d and Conv3d

  #12 opened 3 years ago by swilson314
  1

• Imprecision?

  #11 opened 4 years ago by slerman12
  0

• Getting error from running README example...

  #9 opened 4 years ago by slerman12
  2

• RuntimeError: grad can be implicitly created only for scalar outputs

  #7 opened 4 years ago by slerman12
  1

• RuntimeError: grad can be implicitly created only for scalar outputs

  #8 opened 4 years ago by slerman12
  0

• Can compute_grad1 support multiple backprops?

  #6 opened 4 years ago by zhangxz1123
  0

• A One-Line Bug That Could be Easily Fixed

  #5 opened 5 years ago by Haoxiang-Wang
  1

• Support of padding, dilation and stride in compute_grad1 function

  #3 opened 5 years ago by oljike
  0

• Replace module backward hook with tensor hook

  #1 opened 5 years ago by yaroslavvb
  0