---
license: bsd-3-clause
tags:
- audio-text
- autoaugment
- leave-one-out
- pseudo-label
- random
- security
- standard
- strict
- tfrecord
---

# clean.py

## Dataset Summary

A **security** dataset with **audio text** modality, stored in **tfrecord** format.

## Preprocessing & Augmentation

- **Preprocessing**: standard
- **Augmentation**: autoaugment

## Splits & Sampling

- **Split strategy**: leave one out
- **Sampling**: random

## Quality & Labeling

- **Quality filtering**: strict
- **Labeling**: pseudo label

## Files

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

## License

See the license field above.