Hugging Face's logo

Datasets:
mvp-lab
/
LLaVA-OneVision-2-Data

Tasks:
Video-Text-to-Text
Visual Question Answering
Image-Text-to-Text
Modalities:
Image
Text
Video
Formats:
parquet
Languages:
English
Size:
< 1K
ArXiv:
Tags:
llava
multimodal
video
spatial-reasoning
Libraries:
Datasets
pandas
Polars
License:
Dataset card Data Studio Files
xet
 Community
2
LLaVA-OneVision-2-Data / viewer

Commit History

  1. Simplify spatial parquet schema to id+task+image+preview

    xiangan
    Verified signature
    4509d45

  2. Simplify caption parquet schema to video+preview only

    xiangan
    Verified signature
    2e8a907

  3. Embed video bytes in caption parquet previews (30s/60s/180s/gt10min)

    xiangan
    Verified signature
    6c95777

  4. Add gt10min caption preview parquet

    xiangan
    Verified signature
    525bda9

  5. Add 180s caption preview parquet

    xiangan
    Verified signature
    cab7a7a

  6. Add Parquet preview samples for HF dataset viewer

    xiangan
    Verified signature
    480fd76