vesm_datasets / README.md
tuandinh's picture
Create README.md
6af85f4 verified
|
Raw
History Blame Contribute Delete
773 Bytes
metadata
language:
  - en
tags:
  - biology
  - ESM
  - language-model
  - protein
  - VEP
pretty_name: VESM scores
size_categories:
  - 100M<n<1B

Training and Inference Datasets for VESM

This repository provides datasets for training and evaluating VESM models in the paper "Compressing the collective knowledge of ESM into a single protein language model" by Tuan Dinh, Seon-Kyeong Jang, Noah Zaitlen and Vasilis Ntranos.

  • train.zip: training datasets
  • benchmarks.zip: evaluation datasets
  • base_llrs.zip: individual LLRS used for creating ESMIN dataset.

Please see the corresponding GitHub repo (https://github.com/ntranoslab/vesm) for more details.

License

The predictions of VESM_3B are distributed under the MIT License.