RuntimeError: one of the variables needed for gradient computation has been modified by an inplace operation: [torch.FloatTensor [64, 64, 1, 1]], which is output 0 of ReluBackward0, is at version 1; expected version 0 instead

[KFrank]

Hi Chao!

@Chao_Fan

i am facing this problem and i don’t know how to solve it.

Here is a post that discusses how to diagnose and fix inplace-modification
errors:

"RuntimeError: one of the variables needed for gradient computation has been modified by an inplace operation: [torch.FloatTensor [64, 1]], which is output 0 of AsStridedBackward0, is at version 3; expected version 2 instead. Hint: the backtrace further a

Hi Fahmyadan and Sangyoon! Here are some suggestions about how to track down (and maybe fix) inplace-modification errors. Note that an inplace modification in the forward pass is not necessarily* an error – it depends on whether and how the tensor that was modified is used in the backward pass. Note that inplace operations can be useful for saving memory – if you replace an innocent inplace operation with an out-of-place equivalent, your training will use more memory (and, to a minor e…

Please feel free to post follow-up questions if you’re still having problems
with this issue.

Good luck!

K. Frank