Skip to content

Dataset · BAAI

1M Dataset with Detailed Image Descriptions

BAAI/DenseFusion-1M

A large multimodal set of 1 million to 10 million image-text pairs in English.

What's inside

The dataset from BAAI contains 1 million to 10 million entries including images and English text data. It was created using the Perceptual Fusion approach to generate detailed image descriptions taking fine visual details into account.

What you can do with it

  • Training Multimodal Large Language Models (MLLMs) for detailed image description
  • Fine-tuning Visual Question Answering systems to answer complex photo questions
Tasks
Question answering, Visual question answering
Languages
English
Size
1M–10M rows
Downloads this month
3K
Likes
44

How to load it

from datasets import load_dataset

ds = load_dataset("BAAI/DenseFusion-1M")