文件	最后提交记录	最后更新时间
data	reduce top level clutter: only examples/ and scripts/	4 年前
inverse_folding	Canonicalize all license headers, small README fixes (#529) * notebook link and colab link * add some missing license headers * metamorphosis	3 年前
lm-design	Code release for "Language Models Generalize Beyond Natural Proteins". (#527) Co-authored-by: kwanUm <ori.kabeli@gmail.com> Co-authored-by: Tom Sercu <tom.sercu@gmail.com>	3 年前
protein-programming-language	Make programming language reporting less confusing. 😵‍💫 (#578)	3 年前
variant-prediction	Canonicalize all license headers, small README fixes (#529) * notebook link and colab link * add some missing license headers * metamorphosis	3 年前
README.md	ESM-2 Public Release (#252) Source code and model weights release of ESM-2 Protein Language Model from the paper Language models of protein sequences at the scale of evolution enable accurate structure prediction (Z. Lin, H. Akin, R. Rao, B. Hie, Z. Zhu, W. Lu, A.S. Costa, M. Fazel-Zarandi, T. Sercu, S. Candido, A. Rives, 2022)	4 年前
contact_prediction.ipynb	ESM-2 Public Release (#252) Source code and model weights release of ESM-2 Protein Language Model from the paper Language models of protein sequences at the scale of evolution enable accurate structure prediction (Z. Lin, H. Akin, R. Rao, B. Hie, Z. Zhu, W. Lu, A.S. Costa, M. Fazel-Zarandi, T. Sercu, S. Candido, A. Rives, 2022)	4 年前
esm2_infer_fairscale_fsdp_cpu_offloading.py	Canonicalize all license headers, small README fixes (#529) * notebook link and colab link * add some missing license headers * metamorphosis	3 年前
esm_structural_dataset.ipynb	Fix sequence lengths in README & notebooks (#361)	3 年前
sup_variant_prediction.ipynb	Updating READMEs, tests, and notebooks for dir structure	4 年前

What's in this directory

The notebooks are introduced and summarized in ../README.md
data/some_proteins.fasta and its smaller version, data/few_proteins.fasta are a random selection of UniRef50 sequences used in the second example of ../README.md
data/1a3a_1_A.a3m, data/1xcr_1_A.a3m, data/5ahw_1_A.a3m are MSAs distributed with trRosetta, used in contact_prediction.ipynb
data/P62593.fasta is introduced and used in sup_variant_prediction.ipynb
Example MSAs genereated in the same way as the MSAs used for MSA Transformer pre-training:
- data/UniRef50_E9K9Y4.a3m, data/UniRef50_UPI0003108055.a3m, data/UniRef50_UPI0003674933.a3m, from the same sequences as trRosetta: data/hhblits_uniclust_2017_10_1a3a_1_A.a3m, data/hhblits_uniclust_2017_10_1xcr_1_A.a3m, data/hhblits_uniclust_2017_10_5ahw_1_A.a3m.
- Generated with: hhblits -i UniRef50_$id.fas -oa3m UniRef50_$id.a3m -n 3 -d /uniclust30_2017_10/uniclust30_2017_10.
esm2_infer_fairscale_fsdp_cpu_offloading.py shows how to load the ESM-2 15B model with Fairscale's FSDP's CPU offloading capability on a single GPU