Add initial dataset card for StreamingVLM datasets
#1
by
nielsr
HF Staff
- opened
This PR adds the initial comprehensive dataset card for the datasets associated with the StreamingVLM project.
It includes:
- Relevant metadata (
task_categories,language, and additionaltags). - A summary of the StreamingVLM project and the role of the datasets.
- Links to the paper (https://huggingface.co/papers/2510.09608), the GitHub repository (https://github.com/mit-han-lab/streaming-vlm), and the project demo page (https://streamingvlm.hanlab.ai).
- Descriptions of the key datasets (
Inf-Stream-Train,Live-WhisperX-526K,Inf-Stream-Eval). - Instructions for dataset preparation, extracted directly from the GitHub README, as a "Sample Usage" section.
- The BibTeX citation for the StreamingVLM paper.