Torch version

[hhheeexxxuuu]

Hello, I install the version of torch is 1.7.0, however, there is a mistake: RuntimeError: one of the variables needed for gradient computation has been modified by an inplace operation: [torch.cuda.FloatTensor [1, 512, 4, 4]] is at version 2; expected version 1 instead. Hint: enable anomaly detection to find the operation that failed to compute its gradient, with torch.autograd.set_detect_anomaly(True). I modified inplace = True to inplace = false, the error is still existing. Can you give me some advice?

[geekyutao]

Hello, I install the version of torch is 1.7.0, however, there is a mistake: RuntimeError: one of the variables needed for gradient computation has been modified by an inplace operation: [torch.cuda.FloatTensor [1, 512, 4, 4]] is at version 2; expected version 1 instead. Hint: enable anomaly detection to find the operation that failed to compute its gradient, with torch.autograd.set_detect_anomaly(True). I modified inplace = True to inplace = false, the error is still existing. Can you give me some advice?

Hi, I guess it's the pytorch version issue. In higher versions, it is not allowed shorthands such as "loss += loss1" when calculating losses. Try "loss = loss + loss1" instead of the shorthand. If it still not work, then down-grade the version to 1.1 or 1.2. Thanks.

[hhheeexxxuuu]

I will try, thank you!

[geekyutao]

Hello, I install the version of torch is 1.7.0, however, there is a mistake: RuntimeError: one of the variables needed for gradient computation has been modified by an inplace operation: [torch.cuda.FloatTensor [1, 512, 4, 4]] is at version 2; expected version 1 instead. Hint: enable anomaly detection to find the operation that failed to compute its gradient, with torch.autograd.set_detect_anomaly(True). I modified inplace = True to inplace = false, the error is still existing. Can you give me some advice?

Hi, there is another solution. d_loss should be back-propagated immediately once you have calculated it.