Skip to content

128 instead of 64 #17

@Kunjal1999

Description

@Kunjal1999

Hello,
I am trying to use the framework on a custom dataset. In VAEGAN_ADNI_train, I want to resize the images to 128128128 by setting sp_size=128 instead of sp_size=64.
On running the code for training, for the line mean,logvar,code = E(real_images), I get the following exception:

RuntimeError                              Traceback (most recent call last)
<ipython-input-43-3a5c933a2738> in <cell line: 11>()
     19         #real_array.append(real_images)
     20         z_rand = Variable(torch.randn((_batch_size, latent_dim)),requires_grad=False).cuda()
---> 21         mean,logvar,code = E(real_images)
     22         x_rec = G(code)
     23         x_rand = G(z_rand)

5 frames
/usr/local/lib/python3.10/dist-packages/torch/nn/modules/linear.py in forward(self, input)
    112 
    113     def forward(self, input: Tensor) -> Tensor:
--> 114         return F.linear(input, self.weight, self.bias)
    115 
    116     def extra_repr(self) -> str:

RuntimeError: mat1 and mat2 shapes cannot be multiplied (4x262144 and 32768x2048)

What other changes do I need to make to make the framework accept images of various sizes? (64,128,142,256, etc.)

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions