This project started 3 years ago and since then Python and particularly PyTorch have come a long way. We should consider updating to more modern versions.
We'll probably have to upgrade to get the best sparse matrix support.
I propose we tag a release for the current code, maybe 0.1.1 or something, so we can revert easily if needed.
This project started 3 years ago and since then Python and particularly PyTorch have come a long way. We should consider updating to more modern versions.
We'll probably have to upgrade to get the best sparse matrix support.
I propose we tag a release for the current code, maybe 0.1.1 or something, so we can revert easily if needed.