Make an alternate gpu environment specification

OpenBioML / protein-lm-scaling

Other

54 stars 15 forks source link

Make an alternate gpu environment specification #53

Closed othertea closed 9 months ago

othertea commented 9 months ago

Our current yaml for creating environments is for cpu only and does not specify downloading the cuda version of pytorch. This issue is to create an alternate specification of dependencies for computers with gpus. You can do this by

creating another yaml that specifies alternate dependencies
modifying setup.py
creating an appropriate pyproject.toml

Of course, we are open to other ways of doing this, if you prefer!

Muedi commented 9 months ago

/take

othertea commented 9 months ago

Sorry, to clarify, I meant you could do one of those options, not that you had to do all three!

Muedi commented 9 months ago

I went the yaml way and added an extra one for the cuda install and added the info to the readme. I'll send a PR when my other one is through :)

pascalnotin commented 9 months ago

Issue addressed with PR#50. Thank you @Muedi !