README.md

---
pretty_name: K600 Frozen Test Clips (for GRW Smoothing)
language: en
license: other
task_categories:
  - video-classification
tags:
  - kinetics-600
  - video
  - webdataset
  - evaluation
---

# K600 Frozen Test Clips (for reproducible evaluation)

This repository provides a **frozen** set of extracted clips used for evaluation in the paper:

## Paper
- https://huggingface.co/papers/2511.20928
- https://arxiv.org/abs/2511.20928

## Why this dataset exists
Kinetics videos are hosted on YouTube, and availability can change over time. This dataset exists to provide a **stable evaluation set** aligned with the paper’s experiments.

## What’s included
- `k600_test_ds.tar.gz` — an archive containing extracted test clips organized by class.

## Code (reproduction instructions)
For the **exact, step-by-step instructions** to reproduce results (expected directory layout, caching, and evaluation pipeline), please refer to:
- https://github.com/cmusatyalab/grw-smoothing

## Related model weights
- https://huggingface.co/DrGil/grw-smoothing-movinet

## Licensing / redistribution note
This dataset is marked as `other` because it is derived from Kinetics/YouTube-sourced videos and may involve varying rights. Please ensure your usage complies with the original rights and applicable terms.

## Citation
If you use this dataset, please cite:

```bibtex
@inproceedings{goldman2025grwsmoothing,
  title     = {Smooth Regularization for Efficient Video Recognition},
  author    = {Gil Goldman and Raja Giryes and Mahadev Satyanarayanan},
  booktitle = {Advances in Neural Information Processing Systems (NeurIPS)},
  year      = {2025},
  url       = {https://arxiv.org/abs/2511.20928}
}