Skip to content

Dataset · stepfun-ai

Benchmark for evaluating image editing models

stepfun-ai/GEdit-Bench

A dataset of 1K–10K image and text instruction pairs for testing image editing models on real-world scenarios.

What's inside

Created by the StepFun team for the Step1X-Edit project, this dataset serves as a benchmark for image-to-image tasks. It contains between 1K and 10K records featuring images and text prompts based on real user image editing scenarios.

What you can do with it

  • Evaluating image editing model quality based on text instructions
  • Comparing generative image-to-image models on realistic tasks
Tasks
Image → image
Size
1K–10K rows
Downloads this month
2.5K
Likes
33

How to load it

from datasets import load_dataset

ds = load_dataset("stepfun-ai/GEdit-Bench")