Synthyra/omg_prot50

OMGProt50 with evaluation splits

Thanks Tatta Bio for putting together such an amazing dataset!

To create this version we removed IDs to save space and added the evaluations sets.

See here for a pretokenized version

We add validation and test sets for evalution purposes, including ESM2 speed runs. OMG prot50 was clusterd at 50% identity, so random splits are nonredundant to the training set by default. Random splits of 10,000 make up the base components of the validation and test sets. To the test set, we also add all new Uniprot entries since OMG creation that have transcript level evidence after dedpulication.

Code