Closed rokieplayer20 closed 1 month ago
See #2 , I don't know if you can get embeddings in batches but it works for processing one sequence at a time.
You might be able to use the raw forward
function to accomplish this: https://github.com/evolutionaryscale/esm/blob/main/examples/seqfun_struct.py
Thank you for the great model.
Like ESM2 (https://huggingface.co/facebook/esm2_t6_8M_UR50D), I want to input the protein sequence and turn it into a token and then obtain a tensor (batch size, tokens, channel).
Does the module used like ESM2 have not been released yet? If there is a way, I would appreciate it if you could show me an example.