We read every piece of feedback, and take your input very seriously.
To see all available qualifiers, see our documentation.
Grouped convolution is supported well in pytorch's convolution layers / ops. If possible, it would a great to add that ability to torchsparse.
pytorch
torchsparse