RuntimeError: one of the variables needed for gradient computation has been modified by an inplace operation: [torch.cuda.FloatTensor [2, 256, 25, 36]], which is output 0 of CudnnConvolutionBackward, is at version 1; expected version 0 instead

训练的时候报错：

RuntimeError: one of the variables needed for gradient computation has been modified by an inplace operation: [torch.cuda.FloatTensor [2, 256, 25, 36]], which is output 0 of CudnnConvolutionBackward, is at version 1; expected version 0 instead

主要是梯度求导的问题，可以看下代码里有没有a += xxx这类的代码，把它改成：

a = a + xxx

或者a = a.clone() + xxx

就可以解决问题了。