---
license: apache-2.0
task_categories:
- text-to-video
extra_gated_prompt: "This dataset is curated for the Identity-Preserving Video Generation Challenge (https://hidream-ai.github.io/ipvg-challenge-2026.github.io/), which will be hosted at ACM Multimedia 2026. To request access to this dataset, please complete the registration form (https://forms.gle/j4Nwq38W9TjtPNgq9) using your Hugging Face registered email address. Access requests will be reviewed and processed within 48 hours."
extra_gated_fields:
  I hereby acknowledge and agree that this dataset will be used exclusively for academic research and non-commercial purposes: checkbox

---

# VIP-200K-Video

## Overview

This repository contains **only the video files** (`.tar` archives) from the [VIP-200K](https://huggingface.co/datasets/HiDream-ai/VIP-200K) dataset.

> **For the complete dataset** (including JSON annotations, face frames, and segment metadata), please download [HiDream-ai/VIP-200K](https://huggingface.co/datasets/HiDream-ai/VIP-200K).

## File Structure

```
video/
├── vip200k_train_0001_of_0100.tar
├── vip200k_train_0002_of_0100.tar
├── ...
└── vip200k_train_0100_of_0100.tar
```

Each `.tar` archive contains video clips organized by YouTube video ID:

```
{video_id}/
└── video/
    └── {id}_{beg:05d}_{end:05d}_{fidx_beg:05d}_{fidx_end:05d}.mp4
```

## Usage

Download and extract a specific shard:

```bash
# Download a single shard
huggingface-cli download HiDream-ai/VIP-200K-Video \
    video/vip200k_train_0001_of_0100.tar \
    --repo-type dataset --local-dir .

# Extract
tar -xf vip200k_train_0001_of_0100.tar
```

Download all video shards:

```python
from huggingface_hub import snapshot_download

snapshot_download(
    repo_id="HiDream-ai/VIP-200K-Video",
    repo_type="dataset",
    local_dir="./VIP-200K-Video",
)
```

## Complete Dataset

For annotations (JSON files with face bounding boxes, captions, segment metadata) and face frame images, please refer to the full dataset:

**[HiDream-ai/VIP-200K](https://huggingface.co/datasets/HiDream-ai/VIP-200K)**

## Citation

If you use the VIP-200K dataset or find our research helpful, please cite our paper:

```bibtex
@inproceedings{zhang2025identity,
  title={Identity-Preserving Video Generation Challenge},
  author={Zhang, Yiheng and Qiu, Zhaofan and Cai, Qi and Li, Yehao and Long, Fuchen and Pan, Yingwei and Yao, Ting and Mei, Tao},
  booktitle={Proceedings of the 33rd ACM International Conference on Multimedia},
  pages={13737--13742},
  year={2025}
}
```