datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
gpt-edit-simplerGPT-Image-Edit-1.5M
GPT-Image-Edit-1.5M A Million-Scale, GPT-Generated Image Dataset
📃Arxiv | 🌐 Project Page | 💻Github
GPT-Image-Edit-1.5M is a comprehensive image editing dataset that is built upon HQ-Edit, UltraEdit, OmniEdit and Complex-Edit, with all output images regenerated with GPT-Image-1.
📣 News
[2025.08.20] 🚀 We provide a script for multi-process downloading. See Multi-process Download.
[2025.07.27] 🤗 We release GPT-Image-Edit, a state-of-the-art image editing model with… See the full description on the dataset page: https://huggingface.co/datasets/UCSC-VLAA/GPT-Image-Edit-1.5M.EditReward-CompassHQ-Edit
Dataset Card for HQ-EDIT
HQ-Edit, a high-quality instruction-based image editing dataset with total 197,350 edits. Unlike prior approaches relying on attribute guidance or human feedback on building datasets, we devise a scalable data collection pipeline leveraging advanced foundation models, namely GPT-4V and DALL-E 3.
HQ-Edit’s high-resolution images, rich in detail and accompanied by comprehensive editing prompts, substantially enhance the capabilities of existing image editing… See the full description on the dataset page: https://huggingface.co/datasets/UCSC-VLAA/HQ-Edit.MathCanvas-Edit
MathCanvas-Edit Dataset
🚀 Data Usage
from datasets import load_dataset
dataset = load_dataset("shiwk24/MathCanvas-Edit")
print(dataset)
📖 Overview
MathCanvas-Edit is a large-scale dataset containing 5.2 million step-by-step editing trajectories, forming a crucial component of the [MathCanvas] framework. MathCanvas is designed to endow Unified Large Multimodal Models (LMMs) with intrinsic… See the full description on the dataset page: https://huggingface.co/datasets/shiwk24/MathCanvas-Edit.NHR-Edit
NoHumanRequired (NHR) Dataset for image editing
🌐 NHR Website |
📜 NHR Paper on arXiv |
💻 GitHub Repository |
🤗 NHR-Edit Dataset (part2) |
🤗 BAGEL-NHR-Edit |
❗️ Important: This is Part 1 of the Dataset ❗️
Please be aware that this repository contains the first part of the full NHR-Edit dataset. To have the complete training data, you must also download the second part.
➡️ Click Here to Access Part 2 on… See the full description on the dataset page: https://huggingface.co/datasets/iitolstykh/NHR-Edit.Complex-Edit
Complex-Edit: CoT-Like Instruction Generation for Complexity-Controllable Image Editing Benchmark
📃Arxiv | 🌐Project Page | 💻Github | 📚Dataset | 📄HF Paper
We introduce Complex-Edit, a comprehensive benchmark designed to systematically evaluate instruction-based image editing models across instructions of varying complexity. To develop this benchmark, we harness GPT-4o to automatically collect a diverse set of editing instructions at scale.
Our approach follows a well-structured… See the full description on the dataset page: https://huggingface.co/datasets/UCSC-VLAA/Complex-Edit.edit3d-bench
Dataset Card for edit3d-bench
This is a FiftyOne dataset with 300 samples.
Installation
If you haven't already, install FiftyOne:
pip install -U fiftyone
Usage
import fiftyone as fo
from huggingface_hub import snapshot_download
# Download the dataset snapshot to the current working directory
snapshot_download(
repo_id="Voxel51/edit3d-bench",
local_dir=".",
repo_type="dataset"
)
# Load dataset from current directory using FiftyOne's… See the full description on the dataset page: https://huggingface.co/datasets/Voxel51/edit3d-bench.Edit3D-Bench
Edit3D-Bench
Paper | Project Page | Code
Edit3D-Bench is a benchmark for 3D editing evaluation, introduced in the paper VoxHammer: Training-Free Precise and Coherent 3D Editing in Native 3D Space.
This dataset comprises 100 high-quality 3D models, with 50 selected from Google Scanned Objects (GSO) and 50 from PartObjaverse-Tiny.
For each model, we provide 3 distinct editing prompts. Each prompt is accompanied by a complete set of annotated 3D assets, including
original 3D asset… See the full description on the dataset page: https://huggingface.co/datasets/huanngzh/Edit3D-Bench.multi_reference_image_editing
Multi-Reference Instruction-Based Image Editing Dataset
Overview
This dataset contains 20,000 high-resolution image pairs and multi-modal instructions designed for training advanced image-to-image editing models. It combines two complementary example types: 10,000 reference-grounded edits, where structural or stylistic changes are driven by up to three provided visual reference images, and 10,000 occlusion-based inpainting/outpainting edits, where the model must… See the full description on the dataset page: https://huggingface.co/datasets/molbal/multi_reference_image_editing.mmconflict-editable-values-1k
MMConflict Editable Values 2K
This dataset contains 2,000 source images with visible atomic values for
multimodal conflict research. It has 100 images in each of 20 categories. Every
image comes from a photograph, scan, captured website, software screenshot, or
page of a source document. The dataset does not contain generated images or
project-rendered examples.
Each row records the source, source URL, license, attribution, visible value,
question, and a candidate box around the… See the full description on the dataset page: https://huggingface.co/datasets/shivank21/mmconflict-editable-values-1k.Uni-Edit-Train-Data
Uni-Edit Training Data: Uni-Edit-148k
Project Page | GitHub Repository | Paper
👀 Intro
We introduce Uni-Edit, an intelligent image editing task that serves as the first general task for Unified Multimodal Model (UMM) tuning. Unlike conventional mixed multi-task training that suffers from inherent task conflicts and requires complex multi-stage pipelines, Uni-Edit breaks this paradigm. It achieves true mutual reinforcement by improving image… See the full description on the dataset page: https://huggingface.co/datasets/Uni-Edit/Uni-Edit-Train-Data.EditReward-Bench
News |
Quick Start |
Benchmark Usage |
Citation
EditScore is a series of state-of-the-art open-source reward models (7B–72B) designed to evaluate and enhance instruction-guided image editing.
✨ Highlights
State-of-the-Art Performance: Effectively matches the performance of leading proprietary VLMs. With a self-ensembling strategy, our largest model surpasses even GPT-5 on our comprehensive benchmark… See the full description on the dataset page: https://huggingface.co/datasets/EditScore/EditReward-Bench.emu_edit_test_set
Dataset Card for the Emu Edit Test Set
Dataset Summary
To create a benchmark for image editing we first define seven different categories of potential image editing operations: background alteration (background), comprehensive image changes (global), style alteration (style), object removal (remove), object addition (add), localized modifications (local), and color/texture alterations (texture).
Then, we utilize the diverse set of input images from the MagicBrush… See the full description on the dataset page: https://huggingface.co/datasets/facebook/emu_edit_test_set.NHR-Edit-part2
NoHumanRequired (NHR) Dataset for image editing
🌐 NHR Website |
📜 NHR Paper on arXiv |
💻 GitHub Repository |
🤗 NHR-Edit Dataset (part1) |
🤗 BAGEL-NHR-Edit-V2 |
❗️ Important: This is Part 2 of the Dataset ❗️
Please be aware that this repository contains the second part of the full NHR-Edit dataset. To have the complete training data, you must also download the first part.
➡️ Click Here to Access Part 1 on… See the full description on the dataset page: https://huggingface.co/datasets/iitolstykh/NHR-Edit-part2.VITON-HD-edit
VITON-HD-edit
This repository contains the VITON-HD-edit dataset presented in the paper CtrlVTON: Controllable Virtual Try-On via Visual-Instance-Prompt Segmentation.
Github Repository | Paper (arXiv)
Dataset Overview
The VITON-HD-edit dataset is a public benchmark built to support three evaluations:
Image-editing Virtual Try-On (VTO)
Instance-level visual-prompt segmentation
Spatial controllability
Training the editing model requires triplets $(p… See the full description on the dataset page: https://huggingface.co/datasets/NXN-Labs/VITON-HD-edit.GPT-Image-Edit-1M
GPT-Image-Edit-1M Review Artifact
GPT-Image-Edit-1M is a non-commercial research artifact for instruction-guided image editing. It contains GPT-Image-1 regenerated image-editing triplets, auditable quality-control metadata, and a 200-case human-audit package used to calibrate automated judges in the paper.
License: CC BY-NC-SA 4.0, subject to upstream dataset licenses and applicable third-party service terms.
Reviewer note. The Hugging Face Dataset Viewer shows a 400-row inspection… See the full description on the dataset page: https://huggingface.co/datasets/meimeirun/GPT-Image-Edit-1M.FiVE-Fine-Grained-Video-Editing-Benchmark
FiVE-Bench
FiVE-Bench: A Fine-Grained Video Editing Benchmark for Evaluating Diffusion and Rectified Flow Models
Minghan Li1*, Chenxi Xie2*, Yichen Wu13, Lei Zhang2, Mengyu Wang1†
1Harvard University 2The Hong Kong Polytechnic University 3City University of Hong Kong
*Equal contribution †Corresponding Author
💜 Leaderboard (coming soon) |
💻 GitHub |
🤗 Hugging Face
📝 Project Page |
📰 Paper |
🎥 Video Demo
FiVE is a benchmark comprising 100 videos for… See the full description on the dataset page: https://huggingface.co/datasets/LIMinghan/FiVE-Fine-Grained-Video-Editing-Benchmark.Edit3D-BenchThis repository contains Edit3D-Bench, a 3D editing benchmark proposed in the paper Feedforward 3D Editing via Text-Steerable Image-to-3D.
Project page: https://glab-caltech.github.io/steer3d/
Code: https://github.com/ziqi-ma/Steer3D
The metadata/ directory stores metadata information (the source and target object guid, and editing text) for texture, removal, and addition - each in a separate .csv file.
The data/ directory contains source images (of unedited object), glbs of both source and… See the full description on the dataset page: https://huggingface.co/datasets/ziqima/Edit3D-Bench.EditVerseBench
EditVerse
This repository contains the instruction-based video editing evaluation benchmark for EditVerseBench in paper "EditVerse: A Unified Framework for Editing and Generation via In-Context Learning".
Xuan Ju12, Tianyu Wang1, Yuqian Zhou1, He Zhang1, Qing Liu1, Nanxuan Zhao1, Zhifei Zhang1, Yijun Li1, Yuanhao Cai3, Shaoteng Liu1, Daniil Pakhomov1, Zhe Lin1, Soo Ye Kim1*, Qiang Xu2*
1Adobe Research 2The Chinese University of Hong Kong 3Johns Hopkins University *Corresponding… See the full description on the dataset page: https://huggingface.co/datasets/sooyek/EditVerseBench.ramanv-image-editing
ramanv-image-editing
Image editing dataset for training FLUX.1-Kontext / InstructPix2Pix style models.
Size
592,141 total editing pairs
Sources: ultraedit
Schema
Each shard tar contains {uid}_src.jpg, {uid}_edit.jpg, {uid}_mask.png (where available).
Metadata per record: instruction, prompt, edit_type, caption_before/after, license, sha256.
Licenses
MagicBrush, InstructPix2Pix, Pico-Banana, HumanEdit: CC-BY-4.0
UltraEdit, AnyEdit… See the full description on the dataset page: https://huggingface.co/datasets/lingamvamshikrishnareddy/ramanv-image-editing.DAVIS-Edit
StableV2V: Stablizing Shape Consistency in Video-to-Video Editing
Chang Liu, Rui Li, Kaidong Zhang, Yunwei Lan, Dong Liu
[Paper] / [Project] / [GitHub] / [Models (Huggingface)] / [Models (wisemodel)] / [DAVIS-Edit (wisemodel)] / [Models (ModelScope)] / [DAVIS-Edit (ModelScope)]
HuggingFace repo of the testing benchmark DAVIS-Edit proposed in the paper titled "StableV2V: Stablizing Shape Consistency in Video-to-Video Editing".
Data Structure
We follow the same data… See the full description on the dataset page: https://huggingface.co/datasets/AlonzoLeeeooo/DAVIS-Edit.Outfit_Qwen-Image-Edit-2511_in_Kling
Outfit_Qwen-Image-Edit-2511_in_Kling
Synthetic outfit-swap pairs for Qwen-Image-Edit-2511 SFT (keyframe garment edit),
generated with IDM-VTON as the teacher over VITON-HD.
Batches
Batches are separate directories in this one repo. Every batch uses a distinct
(person, garment) pairing: no person is paired with the garment they already wear,
and no pair is repeated across batches. batch_meta_*.json records the seed and the
dedup counts, pairs_*.txt the exact… See the full description on the dataset page: https://huggingface.co/datasets/lee31221/Outfit_Qwen-Image-Edit-2511_in_Kling.ImagePulseV2-Edit-Structure
ImagePulseV2 Dataset - Image Structure
The ImagePulseV2 dataset is a collection we constructed for training the Diffusion Templates series of models. It comprises multiple subsets generated using models such as Z-Image-Turbo, Qwen-Image, and Qwen-Image-Edit, based on prompts randomly sampled from DiffusionDB.
Open-source code: DiffSynth-Studio
Technical report: arXiv
Project homepage: GitHub
Documentation: English Version, Chinese Version
Online demo: ModelScope Studio
Model… See the full description on the dataset page: https://huggingface.co/datasets/DiffSynth-Studio/ImagePulseV2-Edit-Structure.editscore-rl-train
Introduction
Training data for OmniGen2 Online-RL using EditScore.
Usage
# meta file: rl.jsonl
# images:
cat images_part_* > images.tar.gz && tar -xzvf images.tar.gz
Citation
@article{luo2025editscore,
title={EditScore: Unlocking Online RL for Image Editing via High-Fidelity Reward Modeling},
author={Xin Luo and Jiahao Wang and Chenyuan Wu and Shitao Xiao and Xiyan Jiang and Defu Lian and Jiajun Zhang and Dong Liu and… See the full description on the dataset page: https://huggingface.co/datasets/chibifire/editscore-rl-train.EditBench
EditBench: A Comprehensive Image Editing Dataset
This datasets is included in the work Schedule Your Edit: A Simple yet Effective Diffusion Noise Schedule for Image Editing.
Table of Contents
Introduction
Dataset Overview
Dataset Structure
Editing Types
Annotations and Metadata
Usage Instructions
Evaluation Metrics
Examples
License and Citation
Contact Information
Introduction
The EditBench is a comprehensive image editing dataset designed to… See the full description on the dataset page: https://huggingface.co/datasets/LonelVino/EditBench.Text_Guided_Image_Editing
Dataset Card
Dataset in ImagenHub.
Citation
Please kindly cite our paper if you use our code, data, models or results:
@article{ku2023imagenhub,
title={ImagenHub: Standardizing the evaluation of conditional image generation models},
author={Max Ku and Tianle Li and Kai Zhang and Yujie Lu and Xingyu Fu and Wenwen Zhuang and Wenhu Chen},
journal={arXiv preprint arXiv:2310.01596},
year={2023}
}
premiere-video-editing-trajectories
Creative Video-Editing Computer-Use Trajectories (Preview)
A preview release of computer-use agent trajectories from professional video-editing work in Adobe Premiere Pro (building vertical short-form social reels). Each step pairs a screenshot with a structured action and a first-person thought grounded in the editor's spoken narration as they worked, so the step-level reasoning reflects real human intent rather than a rationale written after the fact.
A sample of the human… See the full description on the dataset page: https://huggingface.co/datasets/contralabs/premiere-video-editing-trajectories.samples-editInter-Edit-Test
Inter-Edit-Test
Official test benchmark release for the CVPR 2026 paper:
Inter-Edit: First Benchmark for Interactive Instruction-Based Image Editing
This repository hosts the public release of Inter-Edit-Test, a human-annotated benchmark for the Interactive Instruction-based Image Editing (I^3E) task.
Each sample contains:
a source image,
a coarse user-style interaction mask,
a concise editing instruction,
and a ground-truth edited image.
To simplify large-scale distribution on… See the full description on the dataset page: https://huggingface.co/datasets/a1557811266/Inter-Edit-Test.