---
license: mit
tags:
- audio-video
- auto-ml
- curriculum
- legal
- none
- semi-auto
- strict
- temporal
- webdataset
---

# clean.py

## Dataset Summary

A **legal** dataset with **audio video** modality, stored in **webdataset** format.

## Preprocessing & Augmentation

- **Preprocessing**: auto ml
- **Augmentation**: none

## Splits & Sampling

- **Split strategy**: temporal
- **Sampling**: curriculum

## Quality & Labeling

- **Quality filtering**: strict
- **Labeling**: semi auto

## Files

- `clean.py` — main artifact of this repository

## License

See the license field above.