ESM-2 implementation?

[Amelie-Schreiber]

Any chance there will be an update to use ESM-2 models such as facebook/esm2_t6_8M_UR50D? This would make the code easier to use locally for people who have hardware constraints, and would likely show improved performance as well using the some of the larger ESM-2 models, no?

[JonasLi-19]

Do you have any idea about how to train this model on my own dataset? I didn't find any script for training

[SanFran-Me]

I also want to use ESM2 to run the model, can I directly change the code to replace ESM1b to facebook/esm2_t33_650M_UR50D? @brianhie