---
dataset_info:
  features:
  - name: sequence
    dtype: string
  splits:
  - name: train
    num_bytes: 21366361216
    num_examples: 100000000
  - name: valid
    num_bytes: 2245865
    num_examples: 10000
  - name: test
    num_bytes: 2217939
    num_examples: 10000
  download_size: 21442257818
  dataset_size: 21370825020
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
  - split: valid
    path: data/valid-*
  - split: test
    path: data/test-*
---

Smaller version of [OMG_prot50](https://huggingface.co/datasets/tattabio/OMG_prot50) without the IDs to save space. There are added validation and test sets as well.