GPT-IMAGE-EDIT-1.5M: A Million-Scale, GPT-Generated Image Dataset

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Wang, Yuhan, Yang, Siwei, Zhao, Bingchen, Zhang, Letian, Liu, Qing, Zhou, Yuyin, Xie, Cihang
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911079858176000
author Wang, Yuhan
Yang, Siwei
Zhao, Bingchen
Zhang, Letian
Liu, Qing
Zhou, Yuyin
Xie, Cihang
author_facet Wang, Yuhan
Yang, Siwei
Zhao, Bingchen
Zhang, Letian
Liu, Qing
Zhou, Yuyin
Xie, Cihang
contents Recent advancements in large multimodal models like GPT-4o have set a new standard for high-fidelity, instruction-guided image editing. However, the proprietary nature of these models and their training data creates a significant barrier for open-source research. To bridge this gap, we introduce GPT-IMAGE-EDIT-1.5M, a publicly available, large-scale image-editing corpus containing more than 1.5 million high-quality triplets (instruction, source image, edited image). We systematically construct this dataset by leveraging the versatile capabilities of GPT-4o to unify and refine three popular image-editing datasets: OmniEdit, HQ-Edit, and UltraEdit. Specifically, our methodology involves 1) regenerating output images to enhance visual quality and instruction alignment, and 2) selectively rewriting prompts to improve semantic clarity. To validate the efficacy of our dataset, we fine-tune advanced open-source models on GPT-IMAGE-EDIT-1.5M. The empirical results are exciting, e.g., the fine-tuned FluxKontext achieves highly competitive performance across a comprehensive suite of benchmarks, including 7.24 on GEdit-EN, 3.80 on ImgEdit-Full, and 8.78 on Complex-Edit, showing stronger instruction following and higher perceptual quality while maintaining identity. These scores markedly exceed all previously published open-source methods and substantially narrow the gap to leading proprietary models. We hope the full release of GPT-IMAGE-EDIT-1.5M can help to catalyze further open research in instruction-guided image editing.
format Preprint
id arxiv_https___arxiv_org_abs_2507_21033
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle GPT-IMAGE-EDIT-1.5M: A Million-Scale, GPT-Generated Image Dataset
Wang, Yuhan
Yang, Siwei
Zhao, Bingchen
Zhang, Letian
Liu, Qing
Zhou, Yuyin
Xie, Cihang
Computer Vision and Pattern Recognition
Recent advancements in large multimodal models like GPT-4o have set a new standard for high-fidelity, instruction-guided image editing. However, the proprietary nature of these models and their training data creates a significant barrier for open-source research. To bridge this gap, we introduce GPT-IMAGE-EDIT-1.5M, a publicly available, large-scale image-editing corpus containing more than 1.5 million high-quality triplets (instruction, source image, edited image). We systematically construct this dataset by leveraging the versatile capabilities of GPT-4o to unify and refine three popular image-editing datasets: OmniEdit, HQ-Edit, and UltraEdit. Specifically, our methodology involves 1) regenerating output images to enhance visual quality and instruction alignment, and 2) selectively rewriting prompts to improve semantic clarity. To validate the efficacy of our dataset, we fine-tune advanced open-source models on GPT-IMAGE-EDIT-1.5M. The empirical results are exciting, e.g., the fine-tuned FluxKontext achieves highly competitive performance across a comprehensive suite of benchmarks, including 7.24 on GEdit-EN, 3.80 on ImgEdit-Full, and 8.78 on Complex-Edit, showing stronger instruction following and higher perceptual quality while maintaining identity. These scores markedly exceed all previously published open-source methods and substantially narrow the gap to leading proprietary models. We hope the full release of GPT-IMAGE-EDIT-1.5M can help to catalyze further open research in instruction-guided image editing.
title GPT-IMAGE-EDIT-1.5M: A Million-Scale, GPT-Generated Image Dataset
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2507.21033