datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
synthetic_data_v5_finegrain_layout_relight_with_our_synthetic_data_coco_l_full_500kfunsd-layoutlmv3Orchestra_Layout_Dataset
Orchestral Score Layout Dataset
Ground-truth annotations for staff layout detection in orchestral music scores, covering two Tchaikovsky symphonies.
Contents
dataset_layout/
├── Tchai_4.pdf # Source PDF — Tchaikovsky Symphony No. 4
├── Tchai_4/ # Page images (PNG, one per page)
├── Tchai_4_csv_gt/ # Per-page CSV ground truth for Tchai_4
│
├── Tchai_6.pdf # Source PDF — Tchaikovsky Symphony No. 6
├── Tchai_6/… See the full description on the dataset page: https://huggingface.co/datasets/BowenC/Orchestra_Layout_Dataset.ndl-layout-dataset
NDL-DocL Kotenseki Layout Dataset (YOLO format)
A YOLO-formatted conversion of the kotenseki (pre-modern Japanese materials, 古典籍資料)
subset of the NDL-DocL dataset published by the National Diet Library of Japan (NDL).
Source dataset: https://github.com/ndl-lab/layout-dataset
Source images: NDL Digital Collections https://dl.ndl.go.jp/
国立国会図書館が公開する NDL-DocL データセットのうち、古典籍資料を
YOLO 形式（Ultralytics 互換）に変換したものです。
This is a modified/derived version. The bounding boxes were converted… See the full description on the dataset page: https://huggingface.co/datasets/nakamura196/ndl-layout-dataset.RealDoc-Bench-Layout
RealDocBench-Layout
A 1,500-page document-layout benchmark for evaluating layout-detection
models on real-world documents. COCO-style annotations across 9 block
classes.
Contents
images/ — 1,500 page images (PNG / JPG / occasional WebP-as-PNG; see Caveats).
annotations/<pageId>.json — per-page COCO files, each with a single
image record, an annotations list, a categories list, and a
page_info block.
manifest.csv — pageId → domain + source URLs. The canonical row… See the full description on the dataset page: https://huggingface.co/datasets/Extend-AI/RealDoc-Bench-Layout.room-layout-planning-curated-v1
Room Layout Planning — curated pilot v1
39 个逐条检查并编写需求的房间布局任务，供实验流程验证与人工抽查。所有最终设计需求均为 AI 编写；没有人工标注或人工复核声明。
Split
条数
独立房屋
几何来源
train
26
26
InstructScene / 3D-FRONT
val_seen
7
7
InstructScene / 3D-FRONT
val_unseen
6
5
M3DLayout / Matterport3D
输入：英文使用需求 + 可用地板多边形 + 4–10 件家具及固定宽深尺寸。输出：所有家具的二维位置与旋转角度。家具清单和尺寸不可修改。参考摆放已通过几何检查，但没有被认证为满足全部语言偏好的标准答案。
下载后打开 review.html 可以逐条浏览需求、尺寸、空房轮廓、参考图和修订理由。原文与 46 条逐条审核记录见 individual_reviews.jsonl，其中 39 条保留、7 条排除。此次规模适合跑通… See the full description on the dataset page: https://huggingface.co/datasets/yfan1997/room-layout-planning-curated-v1.layout_diffusion_hypersimThis repository contains the data for SceneCraft: Layout-Guided 3D Scene Generation.
Project page: https://orangesodahub.github.io/SceneCraft
Code: https://github.com/OrangeSodahub/SceneCraft
layoutbench
LayoutBench
Release of LayoutBench dataset from Diagnostic Benchmark and Iterative Inpainting for Layout-Guided Image Generation (CVPR 2024 Workshop)
See also LayoutBench-COCO for zero-shot evaluation on OOD layouts with real objects.
[Project Page]
[Paper]
Authors:
Jaemin Cho,
Linjie Li,
Zhengyuan Yang,
Zhe Gan,
Lijuan Wang,
Mohit Bansal
Summary
LayoutBench is a diagnostic benchmark that examines layout-guided image generation models on arbitrary, unseen layouts.… See the full description on the dataset page: https://huggingface.co/datasets/j-min/layoutbench.3d_layout_reasoningDataset for MetaSpatial: Reinforcing 3D Spatial Reasoning in VLMs for the Metaverse
Github: https://github.com/PzySeere/MetaSpatial
layout_diffusion_scannetpp_voxel0.2This dataset is used in the paper SceneCraft: Layout-Guided 3D Scene Generation.
File information
The repository contains the following file information:
LayoutOrderingHardLayoutPrompterA collection of datasets used in LayoutPrompter (NeurIPS2023).
Specifically, publaynet and rico are downloaded from LayoutFormer++, posterlayout is downloaded from DS-GAN, and webui is downloaded from Parse-Then-Place.
We sincerely thank them for the great work they do.
realm_layoutlayoutflow-xai
Query-Driven Attribution Framework for Iterative Layout Generation (LayoutFlow XAI)
These are the results produced by the official implementation of the 2026 paper “Query-Driven Attribution Framework for Iterative Layout Generation”, where we apply the framework to the LayoutFlow method and use Integrated Gradients as the attribution method.
To view the results in a more organized way, switch to the “Files and versions” tab.
Code Repository
For the official code… See the full description on the dataset page: https://huggingface.co/datasets/vaarga/layoutflow-xai.LayoutSAM-eval
LayoutSAM-eval Benchmark
Overview
LayoutSAM-Eval is a comprehensive benchmark for evaluating the quality of Layout-to-Image (L2I) generation models. This benchmark assesses L2I generation quality from two perspectives: region-wise quality (spatial and attribute accuracy) and global-wise quality (visual quality and prompt following). It employs the VLM’s visual question answering to evaluate spatial and attribute adherence, and utilizes various metrics including IR score… See the full description on the dataset page: https://huggingface.co/datasets/HuiZhang0812/LayoutSAM-eval.Layout_Synthetic_Datakhmer-newspaper-layout-dataset
Khmer Newspaper Layout Dataset
Dataset Description
This dataset contains Khmer newspaper layouts with annotated regions for document layout analysis and OCR tasks.
Dataset Summary
Total Examples: 9,344 newspaper layouts
Language: Khmer (Cambodian)
Image Format: PNG
Annotations: LabelMe JSON format with bounding boxes and segmentation masks
Features
config_id: Unique identifier for each sample
image: Newspaper layout image (PNG)… See the full description on the dataset page: https://huggingface.co/datasets/jonny122/khmer-newspaper-layout-dataset.hcmus-doc-layout
HCMUS Document-Layout Detection
Fine-grained document-layout object detection on thesis pages: real scanned
HCMUS (Vietnamese) theses plus programmatically generated synthetic pages.
Layout (mirrored under the repo root)
Path
Contents
images/train/shard00 … shard03
26,988 JPEG (20,292 unique pages: 6,696 human-reviewed real + 13,596 synthetic; each real page replayed twice as rep2_* aliases → effective ~1:1 real:synthetic per epoch)
images/val/
916… See the full description on the dataset page: https://huggingface.co/datasets/hydroshiba/hcmus-doc-layout.nutrient-detection-layout
Nutrient extraction dataset
This dataset contains annotated images of nutrition tables. The goal of this dataset was to train a model to extract nutrient values from nutrition tables, as part of the Nutrisight project.
It contains ~3k samples in total (2.8k for training and 199 for testing). For more information about the project, please refer to the nutrisight directory in the openfoodfacts-ai GitHub repository.
The images were collected from the Open Food Facts database, and… See the full description on the dataset page: https://huggingface.co/datasets/openfoodfacts/nutrient-detection-layout.fineweb-vi-pdf-layout
Fineweb VI PDF Layout (v2)
Vietnamese web pages (from HuggingFaceFW/fineweb-2) rendered to PDF, with per-page
layout annotations for VLM training.
Each row = 1 document:
Column
Type
Description
doc_name
string
Document ID (e.g. doc_000)
url
string
Source URL (metadata)
pdf
binary
Rendered document.pdf
layouts_merged
string
All per-page layout JSONs merged: {"num_pages": N, "pages": [...]}
pages
image list
Rendered page images (page_*.png)
source_html
string… See the full description on the dataset page: https://huggingface.co/datasets/dauvannam321/fineweb-vi-pdf-layout.ingredient-detection-layout-dataset
Dataset Card for "ingredient-detection-layout-dataset"
More Information needed
random-layoutsLayoutLM_Training_datasetdream-layout-svg-dataset-v11016 images of Random Boxs Along Bézier Curves
ardb-layout-coco-v1
ARDB Daily Bulletin Layout Detection (COCO, folder format)
⚠️ Please use Soxavin/ardb-layout-coco-v2 instead
This v1 repository is the original folder / COCO export and is no longer maintained.
The identical data is published in Parquet format — with a working Dataset Viewer and the
standard 🤗 datasets loader — at
Soxavin/ardb-layout-coco-v2.
New users should load v2:
from datasets import load_dataset
ds = load_dataset("Soxavin/ardb-layout-coco-v2")
A… See the full description on the dataset page: https://huggingface.co/datasets/Soxavin/ardb-layout-coco-v1.3d_layout_reasoning_800vivid_layoutlayoutlm_resume_datacollage-layout-dataset
Collage Layout Synthetic Dataset
Synthetic photo-collage layouts for layout-quality analysis & correction,
built on a six-ingredient framework (Format, Photos, Visual Weight, Hierarchy,
Readability, Harmony). Corrector-not-generator: every collage carries a
naive (v1_center) and a corrected (fit) placement, so a model can learn the
correction. Faces are synthetically replaced (privacy-safe).
How to load
from datasets import load_dataset
ds =… See the full description on the dataset page: https://huggingface.co/datasets/jjobear/collage-layout-dataset.Layout_Synthetic_v2