RuntimeError: Input type (MPSFloatType) and weight type (torch.FloatTensor) should be the same.

[iam-dante]

I have placed all the weghts and model in the same device still this same error so please dont answer me with .to(device) on anything. And yes am using Mac Silicion Chip M2 pro.

[iam-dante]

So after experiments don't run the summary code

from torchinfo import summary

summary(model=model,
        input_size=(1,3,224,224),
        col_names=["input_size", "output_size", "num_params", "trainable"],
        col_width=20,
        row_settings=["var_names"])

The make the model back to cpu.

[justinkkgoh]

Am running M2 on my mac and skipping the summary also fixed this issue for me. Thanks for sharing

[ringdk]

This was exactly it, thank you @iam-dante