Issue with Inplace Operation during Training

[pantheon5100]

Dear author,

I hope you're doing well. I encountered an error while setting up and training the model. The error occurs during the l_d_real.backward() call, and the traceback points to the following message:

RuntimeError: one of the variables needed for gradient computation has been modified by an inplace operation: [torch.cuda.FloatTensor [256]] is at version 5; expected version 4 instead. Hint: the backtrace further above shows the operation that failed to compute its gradient. The variable in question was changed in there or anywhere later.

It seems that an inplace operation is causing an issue with the gradient computation, but I'm unsure of how to resolve it. Could you kindly help me troubleshoot and resolve this problem?

Thank you very much for your time and assistance!

Best regards,