expanded_attrs = attrs.view(*attrs.shape, 1, 1).expand(*attrs.shape, w, h) seems to be a problem

[SoEzreal]

    if self.n_attrs:
        batch_size, c, w, h = x.shape
        expanded_attrs = attrs.view(
            *attrs.shape, 1, 1).expand(*attrs.shape, w, h)
        x = torch.cat((x, expanded_attrs), dim=1)

---

x = torch.cat((x, expanded_attrs), dim=1)
RuntimeError: Tensors must have same number of dimensions: got 4 and 6

---

Hello, when I run this program （VAEGAN）, these sentences of code will have problems. The original data is 4-dimensional, and the torch cannot be executed after dimension upgrading or expansion torch.cat operation

[SoEzreal]

(img, attrs) in enumerate(data_loader）

Because I'm trying to run vaegan on the cifar-10 dataset. In fact, I don't know what “img” and “attrs” from each batch of CELEBA dataset look like.