Skip to content

Dataset · OneThink

OneThinker-600k multimodal dataset

OneThink/OneThinker-train-data

Large-scale multimodal data corpus for training a universal reasoning model.

What's inside

Contains images, videos, and English texts numbering between 10,000 and 100,000 examples. Designed to train the OneThinker multimodal reasoning model, which supports image and video processing tasks.

What you can do with it

  • Training universal multimodal models (image/video-text-to-text)
  • Object detection and image segmentation tasks
Tasks
Image text → text, Video text → text, Object detection, Image segmentation
Languages
English
Size
10K–100K rows
Downloads this month
24.3K
Likes
14

Models trained on it

How to load it

from datasets import load_dataset

ds = load_dataset("OneThink/OneThinker-train-data")