README.md

---
language:
  - en
configs:
  - config_name: default
    data_files:
      - split: train
        path: train_both.parquet
      - split: val
        path: val_both.parquet
---
only dipromats, downsample P