# muyuho/opcdmini

HF-ready prepared subset for `opcdimage`.

Contents:

- `prepared/train.parquet`
- `prepared/val.parquet`
- `images/original_images/*`
- `images/crop/*`
- `summary.json`

This dataset is the minimal public subset required by the current `opcdimage` training recipe.