pyt-team / TopoModelX

Topological Deep Learning
MIT License
219 stars 79 forks source link

Jude/direct sparse #232

Closed jkhouja closed 11 months ago

jkhouja commented 11 months ago

Here's a WIP PR with proposed solution for #209. I tried few variations and this seems to be the most efficient. It doesn't seem to affect speed on smaller dimensions such as here but definitely on larger sparse matrix.

Since I'm still getting familiar with the repo, I'm not sure where's the best place for this function so would love some feedback.