RuntimeError: Mismatch in shape: grad_output[0] has a shape of torch.Size([1]) and output[0]...

训练tfvaegan时出现以下错误：

Traceback (most recent call last):
  File "D:\Prongs Project\tfvaegan-master\train_images.py", line 172, in <module>
    criticD_real.backward(mone)
  File "D:\Anaconda\envs\pyt11\lib\site-packages\torch\_tensor.py", line 363, in backward
    torch.autograd.backward(self, gradient, retain_graph, create_graph, inputs=inputs)
  File "D:\Anaconda\envs\pyt11\lib\site-packages\torch\autograd\__init__.py", line 166, in backward
    grad_tensors_ = _make_grads(tensors, grad_tensors_, is_grads_batched=False)
  File "D:\Anaconda\envs\pyt11\lib\site-packages\torch\autograd\__init__.py", line 50, in _make_grads
    raise RuntimeError("Mismatch in shape: grad_output["
RuntimeError: Mismatch in shape: grad_output[0] has a shape of torch.Size([1]) and output[0] has a shape of torch.Size([]).

将以下代码

input_res = torch.FloatTensor(opt.batch_size, opt.resSize)
input_att = torch.FloatTensor(opt.batch_size, opt.attSize) #attSize class-embedding size
noise = torch.FloatTensor(opt.batch_size, opt.nz)
one = torch.FloatTensor([1])
mone = one * -1

改成

input_res = torch.FloatTensor(opt.batch_size, opt.resSize)
input_att = torch.FloatTensor(opt.batch_size, opt.attSize) #attSize class-embedding size
noise = torch.FloatTensor(opt.batch_size, opt.nz)
one = torch.tensor(1, dtype=torch.float)
mone = one * -1

后续出现另一个错误

 print('[%d/%d]  Loss_D: %.4f Loss_G: %.4f, Wasserstein_dist:%.4f, vae_loss_seen:%.4f'% (epoch, opt.nepoch, D_cost.data[0], G_cost.data[0], Wasserstein_D.data[0],vae_loss_seen.data[0]),end=" ")
IndexError: invalid index of a 0-dim tensor. Use `tensor.item()` in Python or `tensor.item<T>()` in C++ to convert a 0-dim tensor to a number

修改代码为

print('[%d/%d]  Loss_D: %.4f Loss_G: %.4f, Wasserstein_dist:%.4f, vae_loss_seen:%.4f'% (epoch, opt.nepoch, D_cost.item(), G_cost.item(), Wasserstein_D.item(), vae_loss_seen.item()), end=" ")

问题解决