Skip to content

Dataset · MMInstruction

Multimodal instructions

MMInstruction/M3IT

Multimodal instructions in English and Chinese for training vision-language models; 1 to 10 million examples.

What's inside

A set of multimodal instructions for vision-language models; languages are English and Chinese. According to metadata, the size is from 1 to 10 million examples. Tasks include image-to-text and image-classification; the exact composition and format are not disclosed in the card.

What you can do with it

  • Training VLM to follow instructions
  • LoRA for image captioning
  • Evaluating a bilingual vision-language model
Tasks
Image → text, Image classification
Languages
English, Chinese
Size
1M–10M rows
Downloads this month
9.1K
Likes
136

Models trained on it

How to load it

from datasets import load_dataset

ds = load_dataset("MMInstruction/M3IT")