[error] RuntimeError: one of the variables needed for gradient computation has been modified by an inplace operation: [torch.cuda.FloatTensor [8192, 180]], which is output 0 of ViewBackward, is at version 1; expected version 0 instead. Hint: enable anom..

forward 중에 inplace operaion 이 있어서 역전파 autograd시 무슨 문제가 있다는 말

 

x += model(x) 이런 꼴이 있는지 확인해보자

t = model(x)

x = x+t 

이렇게 바꾸면 됨

 

 

https://discuss.pytorch.org/t/runtimeerror-one-of-the-variables-needed-for-gradient-computation-has-been-modified-by-an-inplace-operation-torch-cuda-floattensor-1-1-256-256-which-is-output-0-of-relubackward0-is-at-version-1-expected-version-0-instead/143490/9

RuntimeError: one of the variables needed for gradient computation has been modified by an inplace operation: [torch.cuda.FloatT