Skip to content

Dataset · deepghs

23M captioned Midjourney images archive

deepghs/midjourney_captioned_23m_full

Huge archive of 23.1M original Midjourney images and their English text descriptions.

What's inside

The dataset contains 23,167,456 original Midjourney images paired with English text descriptions and technical metadata (ID, width, height, etc.). Data is provided by an anonymous third-party source and updated in December 2024.

What you can do with it

  • Training and fine-tuning text-to-image generation models
  • Training multimodal models and encoders like CLIP
  • Creating automated image captioning and prompt generation systems
Tasks
Image classification, Zero shot image classification, Text → image
Languages
English
Size
10M–100M rows
Downloads this month
617
Likes
36
Access
on request on Hugging Face

How to load it

from datasets import load_dataset

ds = load_dataset("deepghs/midjourney_captioned_23m_full")