datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
Total_Editing_Synthetic_Video_Albedo_Fullramanv-image-editing-syntheticmulti_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.ramanv-image-editing-pairsidentity_preservation_image_editing
Identity Preservation Augmentation Dataset for Image Editing
Overview
This dataset contains algorithmically generated image pairs designed to
teach diffusion-based image editing models pixel-level identity
preservation — the ability to keep unchanged regions of an image exactly
intact while applying targeted edits.
Every example consists of a reference image, a target image, and a short
natural-language prompt. The transformation between reference and target is… See the full description on the dataset page: https://huggingface.co/datasets/molbal/identity_preservation_image_editing.Model_Editing_HurtHalluEditBench
Can Knowledge Editing Really Correct Hallucinations?
Respository Oveview: This repository contains the code, results and dataset for the paper "Can Knowledge Editing Really Correct Hallucinations? (ICLR 2025)"
TLDR: We proposed HalluEditBench to holistically benchmark knowledge editing methods in correcting real-world hallucinations on five dimensions including Efficacy, Generalization, Portability, Locality, and Robustness. We find that their effectiveness could be far from what… See the full description on the dataset page: https://huggingface.co/datasets/llm-editing/HalluEditBench.SWE-Fixer-Train-Editing-CoT-70KFiVE-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.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.escher-ss2
Dataset Card for escher-ss2
SomethingSomethingv2 dataset
Dataset Structure
Data Instances
Each instance contains:
source_image: The original image
edited_image: The edited version of the image
edit_instruction: The instruction used to edit the image
source_image_caption: Caption for the source image
target_image_caption: Caption for the edited image
Additional metadata fields
Data Splits
{}
escher-aurora-kubric
Dataset Card for escher-aurora-kubric
Aurora-Kubric dataset
Dataset Structure
Data Instances
Each instance contains:
source_image: The original image
edited_image: The edited version of the image
edit_instruction: The instruction used to edit the image
source_image_caption: Caption for the source image
target_image_caption: Caption for the edited image
Additional metadata fields
Data Splits
{}
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}
}
au-editingescher-aurora-ag
Dataset Card for escher-aurora-ag
Aurora-AG dataset
Dataset Structure
Data Instances
Each instance contains:
source_image: The original image
edited_image: The edited version of the image
edit_instruction: The instruction used to edit the image
source_image_caption: Caption for the source image
target_image_caption: Caption for the edited image
Additional metadata fields
Data Splits
{}
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.escher-vismin
Dataset Card for escher-vismin
Vismin dataset
Dataset Structure
Data Instances
Each instance contains:
source_image: The original image
edited_image: The edited version of the image
edit_instruction: The instruction used to edit the image
source_image_caption: Caption for the source image
target_image_caption: Caption for the edited image
Additional metadata fields
Data Splits
{}
escher-human-edit
Dataset Card for escher-human-edit
Human Edit dataset
Dataset Structure
Data Instances
Each instance contains:
source_image: The original image
edited_image: The edited version of the image
edit_instruction: The instruction used to edit the image
source_image_caption: Caption for the source image
target_image_caption: Caption for the edited image
Additional metadata fields
Data Splits
{}
handy-dictation-editing
Handy dictation-editing corpus
Turns a raw dictated transcript into the text the speaker meant to write.
in : um so the meeting is uh moved to friday no wait thursday at three
out: The meeting is Thursday at three.
Three jobs at once, because they are not separable in speech: drop filler words,
repair punctuation and capitalisation, and — the hard one — when the speaker
changes their mind mid-sentence, delete the wording they abandoned and keep only
what they settled on.
Built… See the full description on the dataset page: https://huggingface.co/datasets/MagicNoThief/handy-dictation-editing.escher-magicbrush
Dataset Card for escher-magicbrush
MagicBrush dataset
Dataset Structure
Data Instances
Each instance contains:
source_image: The original image
edited_image: The edited version of the image
edit_instruction: The instruction used to edit the image
source_image_caption: Caption for the source image
target_image_caption: Caption for the edited image
Additional metadata fields
Data Splits
{}
Mask_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}
}
STXBP1_Base_Editing_Parameter_Sweep_V2
STXBP1 Base-Editing Parameter Sweep — v2 (Canonical 603-aa frame)
TL;DR
170 pathogenic STXBP1 variants × 3,850,560 base-editing parameter combinations
per variant = 654.6 million rows. Built on the MANE Plus Clinical canonical
reference NM_003165.6 → NP_003156.1 (603 aa), the same frame
ARIA and ClinVar use.
This is the canonical successor to
SkyWhal3/STXBP1_Base_Editing_Parameter_Sweep
(v1), which was built on NM_001032221.6 (594 aa MANE Select). For variants at
protein… See the full description on the dataset page: https://huggingface.co/datasets/SkyWhal3/STXBP1_Base_Editing_Parameter_Sweep_V2.descript-video-editing-trajectories
Descript Video-Editing Computer-Use Trajectories (Preview)
This is a preview release of computer-use trajectories from experienced video editors working through client-style editing briefs in Descript: cutting vertical short-form social reels from source footage. Each session is a long edit, about two hours and a few hundred steps, and the editor's spoken narration was recorded while they worked and used to ground the step-level reasoning. Most open GUI-agent datasets cover… See the full description on the dataset page: https://huggingface.co/datasets/contralabs/descript-video-editing-trajectories.ramanv-image-editing-assembledramanv-image-editing-2pb-image-editing-10k-sftImage-Gen-or-Image-Editing
Image Gen or Image Editing
This dataset is designed for text classification of prompts provided by users. It determines whether a prompt is intended for image generation or image editing.
precise_benchmark_for_object_level_image_editing
VOCEdits: A benchmark for precise geometric object-level editing
Sample format: (input image, edit prompt, input mask, ground-truth output mask, ...)
Please refer to our paper for more details: "📜 POEM: Precise Object-level Editing via MLLM control", SCIA 2025.
How to Evaluate?
Before evaluation, you should first generate your edited images.
Use datasets library to download dataset. You should only use input image, edit prompt, and id columns to generate edited images.… See the full description on the dataset page: https://huggingface.co/datasets/monurcan/precise_benchmark_for_object_level_image_editing.pb-image-editing-10k-sftSubject_Driven_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}
}