pbwpbw/tiny_llavavideo

---
license: apache-2.0
task_categories:
- video-text-to-text
---

**<center><span style="font-size:2em;">TinyLLaVA-Video</span></center>**

[](https://arxiv.org/abs/2501.15513)[](https://github.com/ZhangXJ199/TinyLLaVA-Video)[](https://huggingface.co/papers/2501.15513)

This dataset combines data from multiple sources for pre-training and fine-tuning.

**Pretrain Data:** Four subsets of LLaVA-Video-178K (`0_30_s_academic_v0_1`, `30_60_s_academic_v0_1`, `0_30_s_youtube_v0_1`, `30_60_s_youtube_v0_1`), supplemented with filtered Video-LLaVA data ([https://huggingface.co/datasets/LanguageBind/Video-LLaVA](https://huggingface.co/datasets/LanguageBind/Video-LLaVA)) and data from Valley ([https://github.com/RupertLuo/Valley](https://github.com/RupertLuo/Valley)). The video data can be downloaded from the linked datasets, and cleaned annotations are provided within this dataset.

**Finetune Data:** Four subsets of LLaVA-Video-178K (`0_30_s_academic_v0_1`, `30_60_s_academic_v0_1`, `0_30_s_youtube_v0_1`, `30_60_s_youtube_v0_1`). Cleaned annotations are provided; video data is available via the LLaVA-Video-178K dataset ([https://huggingface.co/datasets/lmms-lab/LLaVA-Video-178K](https://huggingface.co/datasets/lmms-lab/LLaVA-Video-178K)).

The data is organized as follows:

```Shell
dataset
├── academic_source
├── liwei_youtube_videos
├── valley
├── text_files
│ ├── cleaned_video_caption.json
│ ├── cleaned_video_openqa.json
```

**Note:** If there is any infringement, please contact us for removal. Please refer to the Github repository for detailed instructions on data usage and training.