HuggingFaceM4/FineVisionMax

Name: HuggingFaceM4/FineVisionMax
Creator: HuggingFaceM4
License: Unknown
Keywords: huggingface, task_categories:image-text-to-text, language:en, language:zh, size_categories:10M<n<100M, format:parquet, format:optimized-parquet, modality:image, modality:text, image-text-to-text

Image Text To Text · HuggingFaceM4· 50.2K

Unknown 4.4 TB task_categories:image-text-to-textlanguage:enlanguage:zhsize_categories:10M<n<100Mformat:parquet

FineVision is a massive collection of datasets with 17.3M images, 24.3M samples, 88.9M turns, and 9.5B answer tokens, designed for training state-of-the-art open Vision-Language-Models.

Open in MLForge Sign up free Desktop app

# download instantly
mlforge datasets pull HuggingFaceM4/FineVisionMax

Dataset details

Task

Image Text To Text

Language

License

Unknown

Size

4.4 TB

Rows / images

24.2M

Creator

HuggingFaceM4

Downloads

50.2K

Source

huggingface_datasets

Updated

2025-10-21

About HuggingFaceM4/FineVisionMax

FineVision is a massive collection of datasets with 17.3M images, 24.3M samples, 88.9M turns, and 9.5B answer tokens, designed for training state-of-the-art open Vision-Language-Models.