---
license: mit
tags:
- adaptive
- comics
- domain-specific
- kfold-5
- light
- parquet
- stratified
- video-text
- weak-supervision
---

# dataset_108446237_comics_video_text.py

## Dataset Summary

A **comics** dataset with **video text** modality, stored in **parquet** format.

## Preprocessing & Augmentation

- **Preprocessing**: domain specific
- **Augmentation**: light

## Splits & Sampling

- **Split strategy**: kfold 5
- **Sampling**: stratified

## Quality & Labeling

- **Quality filtering**: adaptive
- **Labeling**: weak supervision

## Files

- `dataset_108446237_comics_video_text.py` — main artifact of this repository

## License

See the license field above.