---
license: cc-by-4.0
tags:
- active
- audio-text
- auto-ml
- heavy
- moderate
- parquet
- self-training
- stratified-90-10
- traffic
---

# preprocess.py

## Dataset Summary

A **traffic** dataset with **audio text** modality, stored in **parquet** format.

## Preprocessing & Augmentation

- **Preprocessing**: auto ml
- **Augmentation**: heavy

## Splits & Sampling

- **Split strategy**: stratified 90 10
- **Sampling**: active

## Quality & Labeling

- **Quality filtering**: moderate
- **Labeling**: self training

## Files

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

## License

See the license field above.