我有一個火炬張量z，我想應用一個變換矩陣，mat并使z輸出的大小與 完全相同z。這是我正在運行的代碼：def trans(z):    print(z)    mat = transforms.Compose([transforms.ToPILImage(),transforms.RandomRotation(90),transforms.ToTensor()])    z = Variable(mat(z.cpu()).cuda())    z = nnf.interpolate(z, size=(28, 28), mode='linear', align_corners=False)    return zz = trans(z)但是，我收到此錯誤：RuntimeError                              Traceback (most recent call last)<ipython-input-12-e2fc36889ba5> in <module>()      3 inputs,targs=next(iter(tst_loader))      4 recon, mean, var = vae.predict(model, inputs[img_idx])----> 5 out = vae.generate(model, mean, var)4 frames/content/vae.py in generate(model, mean, var)     90     z = trans(z)     91     z = Variable(z.cpu().cuda())---> 92     out = model.decode(z)     93     return out.data.cpu()     94 /content/vae.py in decode(self, z)     56      57     def decode(self, z):---> 58         out = self.z_develop(z)     59         out = out.view(z.size(0), 64, self.z_dim, self.z_dim)     60         out = self.decoder(out)/usr/local/lib/python3.6/dist-packages/torch/nn/modules/module.py in _call_impl(self, *input, **kwargs)    720             result = self._slow_forward(*input, **kwargs)    721         else:--> 722             result = self.forward(*input, **kwargs)    723         for hook in itertools.chain(    724                 _global_forward_hooks.values(),/usr/local/lib/python3.6/dist-packages/torch/nn/modules/linear.py in forward(self, input)     89      90     def forward(self, input: Tensor) -> Tensor:---> 91         return F.linear(input, self.weight, self.bias)     92      93     def extra_repr(self) -> str:/usr/local/lib/python3.6/dist-packages/torch/nn/functional.py in linear(input, weight, bias)   1674         ret = torch.addmm(bias, input, weight.t())   1675     else:-> 1676         output = input.matmul(weight.t())   1677         if bias is not None:   1678             output += biasRuntimeError: mat1 dim 1 must match mat2 dim 0如何成功應用此旋轉(zhuǎn)變換mat并且不會出現(xiàn)任何錯誤？