masahi commented on pull request #7441: URL: https://github.com/apache/tvm/pull/7441#issuecomment-780305205
Yes, it's possible but a bit complicated. PyTorch also has `return_counts` option https://pytorch.org/docs/stable/generated/torch.unique.html I think for the first PR, not all options need to be implemented. We can follow up later. I'm using PyTorch GPU impl as reference, see for example below on how they support count https://github.com/pytorch/pytorch/blob/22a34bcf4e5eaa348f0117c414c3dd760ec64b13/aten/src/ATen/native/cuda/Unique.cu#L60-L68 ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [email protected]
