x = torch.randn(50) # create a rank 1 tensor (vector) with 50 features x.shape ... losses.append(float(loss)/batch_size) # compute *average* loss.
確定！ 回上一頁
