fix: make `pytorch_ops` work on GPU

google-research / fast-soft-sort

Fast Differentiable Sorting and Ranking

Apache License 2.0

570 stars 47 forks source link

fix: make `pytorch_ops` work on GPU #23

Open Markus28 opened 9 months ago

Markus28 commented 9 months ago

This PR adds a small modification to the NumpyOpWrapper that is used in pytorch_ops, making the torch functions work with CUDA tensors. This should fix the following two issues:

https://github.com/google-research/fast-soft-sort/issues/16
https://github.com/google-research/fast-soft-sort/issues/18

We have to move the tensors from GPU to CPU and back to run the soft sorting and ranking in numpy. There's no way around that unfortunately.

google-cla[bot] commented 9 months ago

Thanks for your pull request! It looks like this may be your first contribution to a Google open source project. Before we can look at your pull request, you'll need to sign a Contributor License Agreement (CLA).

View this failed invocation of the CLA check for more information.

For the most up to date status, view the checks section at the bottom of the pull request.