Skip to content

Dataset · iitolstykh

First part of the image editing dataset

iitolstykh/NHR-Edit

Synthetic dataset of 100k-1M image-instruction pairs in English for training image editing models.

What's inside

The repository hosts the first part of the NoHumanRequired (NHR-Edit) dataset with 100,000 to 1,000,000 records. Contains images and English text instructions for image-to-image and text-to-image tasks. Authors note that the second part (part2) must be downloaded separately for the full dataset.

What you can do with it

  • Training and fine-tuning text-guided image editing models
  • Training LoRA for diffusion models on image-to-image tasks
  • Evaluating text instruction adherence when modifying images
Tasks
Image → image, Text → image
Languages
English
Size
100K–1M rows
Downloads this month
4.4K
Likes
48

How to load it

from datasets import load_dataset

ds = load_dataset("iitolstykh/NHR-Edit")