why not use built-in "torch.nn.functional.conv2d" for ska?

[DLeeeeeee]

Many thanks for your excellent work and for sharing it with the community.

I noticed that the SKA module includes custom forward and backward functions. Please correct me if I’m mistaken, but SKA seems to be essentially a convolution with dynamic kernel weights. I’m curious—what’s the reason for implementing it manually instead of using PyTorch’s built-in Conv2d? Using the built-in function could potentially simplify deployment.

[Intelligent-Monster]

The same question as above. The SKA module will have problems when converted to the TensorRT format, which is not compatible enough and not conducive to deployment.

[DLeeeeeee]

I guess I found the reason. That is because SKA is not a standard conv. it takes in kernel weights in shape (B, C, K**2, H, W), normal conv kernel is like (in, out, K, K).

[OneWalkman]

Hey, guys, have you successfully completed the deployment testing? I'm encountering a compilation failure with the MMVC component, which is really frustrating. I'd like to ask how you handled it.