search

facebookresearch / esm

Evolutionary Scale Modeling (esm): Pretrained language models for proteins
MIT License
2.97k stars 586 forks source link

confused about the model used #570

Open daisykuma22 opened 1 year ago

daisykuma22 commented 1 year ago

NOTE: if this is not a bug report, please use the GitHub Discussions for support questions (How do I do X?), feature requests, ideas, showcasing new applications, etc. hello, i have used esmfold for protein structure prediction and everything goes well. The command i used was as follows:

esm-fold -i 791.fasta -o ./ -m /mnt/esm-main/esmfold_data/ --chunk-size 32

23/06/06 10:43:11 | INFO | root | Reading sequences from 791.fasta
23/06/06 10:43:11 | INFO | root | Loaded 7914 sequences from 791.fasta
23/06/06 10:43:11 | INFO | root | Loading model
23/06/06 10:43:12 | INFO | torch.distributed.nn.jit.instantiator | Created a temporary directory at /tmp/tmp9xs88ajp
23/06/06 10:43:12 | INFO | torch.distributed.nn.jit.instantiator | Writing /tmp/tmp9xs88ajp/_remote_module_non_scriptable.py
23/06/06 10:44:16 | INFO | root | Starting Predictions

Pretrained ESM data directory list:

esm2_t36_3B_UR50D-contact-regression.pt 
esm2_t36_3B_UR50D.pt 
esmfold_3B_v1.pt

The confusion is that i have no idea which model is used to predict the structure. Can you help with this question？Thanks a lot.

Bug description Please enter a clear and concise description of what the bug is.

Reproduction steps Enter steps to reproduce the behavior.

Expected behavior Give a clear and concise description of what you expected to happen.

Logs Please paste the command line output:

Output goes here

Additional context Add any other context about the problem here. (like proxy settings, network setup, overall goals, etc.)

mdlakic commented 1 year ago

Only one of those checkpoints is called esmfold. The rest of them are for sequence embedding and contact prediction