Gabriele Corso

icon indicating a website link gcorso.github.io icon indicating an email [email protected]

icon company MIT icon location PhD student @ MIT • Research on Machine Learning and Geometric Representation Learning • Previously intern @ Twitter Research, D.E. Shaw and IBM

Results 65 comments of Gabriele Corso

The problem of strange ligand conformations predicted by Boltz-2 through SMILES format.

Did this happen when using or not using the --use_potentials flag?

Meaning of affinity_pred_value

Hi @morgannprice , no the documentation is correct, but the affinity_pred_value should only be used when comparing different binders (e.g. in a hit / lead optimization setting) not when comparing...

Fixed --write_full_pae and --write_full_pde options. Fixes #585

Thank you for pointing this out @tomgoddard ! Could you add a note about --no_write_full_pae/pde in the documentation table?

Specify which device to run on

Hi Diego, thank you for suggesting the change! Could you add some documentation to the inference docs to explain the updated usage of the parameter?

Problem with the triton package when calling boltz predict

Hi @wyb63136 @xavgit @coreyhowe999 @Shredderroy @xavierholt, The solution here should be to add the flag `--no_kernels` to the prediction. The issue occurs in older NVIDIA devices that do not support...