修改网络时
pytroch报错:
RuntimeError: one of the variables needed for gradient computation has been modified by an inplace operation: [torch.cuda.FloatTensor [1600, 16, 256]], which is output 0 of CudnnConvolutionBackward, is at version 1; expected version 0 instead
解决方案一:检查被报错的那个tensor的所有 *** 作,如果出现了x += m
,x=x+m
的加法或者减法运算,全部改为以下格式:
x = x.clone() + m
欢迎分享,转载请注明来源:内存溢出
评论列表(0条)