Michael Plainer comments

Repositories
Issues
Comments

Results 24 comments of


                                            Michael Plainer

Question about Training Data Cutoff Date for Model Checkpoints

Hey, Thank you for your question. We have used only the proteins present in PDBBind v2020. So this should be the cutoff.

Cannot have more than 999 residues

Thank you for posting this issue. Would you have the full stack trace? I haven't worked on this codebase for quite a while.

How to handle platform variants like cuda in workspaces

I would like to chip in here because this is something I (and a lot of other data scientists/researchers) need. As @PhilipVinc already said, a typical use case is to...

How to handle platform variants like cuda in workspaces

A very rudimentary version would be to have the following dependencies based on [PEP508](https://peps.python.org/pep-0508/) ```toml dependencies = [ "jax; platform_system == 'Darwin'", "jax[cuda]; platform_system != 'Darwin'", ] ``` But this...