datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
LayeredFlow-Syn
LayeredFlow-Syn Extracted Ground Truth
This repository contains extracted LayeredFlow ground-truth annotations in Parquet format.
Layout
data/<scene>/<sample>.parquet
For example:
data/0/0_0.parquet
Each Parquet file stores one extracted sample. Rows correspond to files from the extracted
sample directory and include a leftmost visualization image preview when available,
relative_path, num_bytes, sha256, and binary content.
Rows are ordered by frame0/left… See the full description on the dataset page: https://huggingface.co/datasets/princeton-vl/LayeredFlow-Syn.Vera-Layered-Video-Dataset
Dataset for Vera: A Layered Diffusion Model for Content-Preserving Video Editing
Hongkai Zheng¹²* ·
Ta-Ying Cheng² ·
Benjamin Klein² ·
Yisong Yue¹ ·
Zhuoning Yuan²†
¹California Institute of Technology ²Netflix, Inc.
*Work done during an internship at Netflix †Project Lead
TL;DR: A layered diffusion framework for video editing. Vera jointly generates an edit layer, an alpha… See the full description on the dataset page: https://huggingface.co/datasets/netflix/Vera-Layered-Video-Dataset.Transmem_ecsd_qwen3_4b_hotpotqa_layered8_legacyclearwrist-pediatric-wrist-xrayClearWrist: Pediatric Wrist Fracture X-Ray Dataset
20,327 labeled pediatric wrist radiographs, rebuilt from GRAZPEDWRI-DX with clean patient-level splits, verified fracture ground truth, and YOLO-style bounding box annotations.
Overview
This dataset packages the full GRAZPEDWRI-DX corpus, 20,327 pediatric wrist radiographs from 6,091 patients treated at the Department for Pediatric Surgery of the University Hospital Graz between… See the full description on the dataset page: https://huggingface.co/datasets/Layered-Labs/clearwrist-pediatric-wrist-xray.LayeredDepthLayeredDepth is a real-world benchmark for Seeing and Seeing Through the Glass: Real and Synthetic Data for Multi-Layer Depth Estimation . If this dataset was helpful for your research, please cite
@misc{wen2025seeingseeingglassreal,
title={Seeing and Seeing Through the Glass: Real and Synthetic Data for Multi-Layer Depth Estimation},
author={Hongyu Wen and Yiming Zuo and Venkat Subramanian and Patrick Chen and Jia Deng},
year={2025},
eprint={2503.11633}… See the full description on the dataset page: https://huggingface.co/datasets/princeton-vl/LayeredDepth.LayeredDepth-SynLayeredDepth-Syn is a synthetic dataset for Seeing and Seeing Through the Glass: Real and Synthetic Data for Multi-Layer Depth Estimation . If this dataset was helpful for your research, please cite
@misc{wen2025seeingseeingglassreal,
title={Seeing and Seeing Through the Glass: Real and Synthetic Data for Multi-Layer Depth Estimation},
author={Hongyu Wen and Yiming Zuo and Venkat Subramanian and Patrick Chen and Jia Deng},
year={2025},
eprint={2503.11633}… See the full description on the dataset page: https://huggingface.co/datasets/princeton-vl/LayeredDepth-Syn.LayeredFlowneiss-injury-dataNEISS Injury Data
7.3 million emergency department injury records from 2005 to 2024, consolidated into a single query-ready Parquet file.
Overview
NEISS Injury Data is the most complete publicly available consolidation of National Electronic
Injury Surveillance System records, covering 7,326,429 emergency department visits associated
with consumer products, sports, and activities from 2005 through 2024. Each record is one ED
visit drawn… See the full description on the dataset page: https://huggingface.co/datasets/Layered-Labs/neiss-injury-data.ACNE04ACNE04
1,457 facial acne photographs, rebuilt from the original ACNE04 release with clean image/label separation, verified 4-level Hayashi severity grades, and 18,983 lesion-level bounding box annotations.
Overview
This dataset packages the full ACNE04 corpus, 1,457 facial photographs graded for acne severity on the 4-level Hayashi scale, as a single, ready-to-load Hugging Face dataset. Every image is embedded directly in the… See the full description on the dataset page: https://huggingface.co/datasets/Layered-Labs/ACNE04.OmniPSD_Layered_PosterVera-Layered-Video-Dataset
Dataset for Vera: A Layered Diffusion Model for Content-Preserving Video Editing
Hongkai Zheng¹²* ·
Ta-Ying Cheng² ·
Benjamin Klein² ·
Yisong Yue¹ ·
Zhuoning Yuan²†
¹California Institute of Technology ²Netflix, Inc.
*Work done during an internship at Netflix †Project Lead
TL;DR: A layered diffusion framework for video editing. Vera jointly generates an edit layer, an alpha… See the full description on the dataset page: https://huggingface.co/datasets/sky-meilin/Vera-Layered-Video-Dataset.claude-fable-derm
Claude Fable Derm
Claude Fable Derm is a dataset of patient dermatology questions and their raw, unprompted
responses from Claude Fable. Each question is generated by crossing a real dermatological
topic with one of Paul Ekman's six basic emotions, producing emotionally distinct framings
of the same underlying medical concern. Answers are collected with no system prompt or
role instruction, capturing how the model responds to a patient question exactly as it
would in the wild.… See the full description on the dataset page: https://huggingface.co/datasets/Layered-Labs/claude-fable-derm.LayeredDepthLayeredDepth is a real-world benchmark for Seeing and Seeing Through the Glass: Real and Synthetic Data for Multi-Layer Depth Estimation . If this dataset was helpful for your research, please cite
@misc{wen2025seeingseeingglassreal,
title={Seeing and Seeing Through the Glass: Real and Synthetic Data for Multi-Layer Depth Estimation},
author={Hongyu Wen and Yiming Zuo and Venkat Subramanian and Patrick Chen and Jia Deng},
year={2025},
eprint={2503.11633}… See the full description on the dataset page: https://huggingface.co/datasets/PItchayapar/LayeredDepth.Vera-Layered-Video-Dataset
Dataset for Vera: A Layered Diffusion Model for Content-Preserving Video Editing
Hongkai Zheng¹²* ·
Ta-Ying Cheng² ·
Benjamin Klein² ·
Yisong Yue² ·
Zhuoning Yuan²†
¹California Institute of Technology ²Netflix, Inc.
*Work done during an internship at Netflix †Project Lead
TL;DR: A layered diffusion framework for video editing. Vera jointly generates an edit layer, an alpha… See the full description on the dataset page: https://huggingface.co/datasets/3AFamily/Vera-Layered-Video-Dataset.benchbase-medqaBenchBase MedQA
MedQA in the BenchBase unified schema: 11,470 USMLE-style 4-option MCQs for evaluating medical language models.
Overview
BenchBase MedQA is the MedQA benchmark converted into the BenchBase unified schema.
Each item is a USMLE-style 4-option multiple-choice question covering clinical medicine,
pharmacology, pathology, and basic sciences. The unified format standardizes question
structure, answer representation… See the full description on the dataset page: https://huggingface.co/datasets/Layered-Labs/benchbase-medqa.benchbase-medmcqaBenchBase: MedMCQA
193,000+ multiple-choice medical questions from Indian postgraduate medical entrance exams, spanning 2,400 healthcare topics.
Overview
BenchBase: MedMCQA is a repackaging of the MedMCQA dataset for the Layered Labs BenchBase
benchmark suite. The source dataset contains 193,155 multiple-choice questions drawn from
AIIMS and PGMR medical entrance examinations, covering 2,400+ healthcare topics across
clinical medicine… See the full description on the dataset page: https://huggingface.co/datasets/Layered-Labs/benchbase-medmcqa.MAUVIKAT_Product_Posters_Qwen_Image_layered_videos
MAUVIKAT (喵维卡特)_AD_Posters
nyc-clinic-ai-infrastructureNYC Clinic AI Infrastructure
Broadband, electricity, and grid reliability data for all 311 NYC ZIP codes: can a clinic run local AI?
Overview
NYC Clinic AI Infrastructure maps three infrastructure prerequisites for on-premise AI deployment
across all 311 New York City ZIP codes. Each row covers one ZIP code with fixed broadband
subscription rates from the Census ACS, ISP coverage and max speeds from the FCC National
Broadband Map… See the full description on the dataset page: https://huggingface.co/datasets/Layered-Labs/nyc-clinic-ai-infrastructure.open-sora-mini-hf-layered
Open-Sora Mini HF (Layered Format)
Converted from wangxingjun778/open-sora-mini-hf
Structure
├── videos/ # Video-only TAR shards (immutable)
│ ├── mixkit_shard_0000.tar
│ └── ...
├── annotations/ # Annotations (can add new versions)
│ └── captions_v1.parquet # Vision-generated captions
└── manifest.parquet # Index: sample_id -> shard mapping
Stats
Metric
Value
Total samples with captions
8,247… See the full description on the dataset page: https://huggingface.co/datasets/zengxianyu/open-sora-mini-hf-layered.qwen_image_layered_editing_with_qwen_image_edit_magic_brushbenchbase-pubmedqa
BenchBase PubMedQA
PubMedQA normalized into the BenchBase unified schema. Every BenchBase dataset shares the same
columns, so one loader and one evaluation loop cover all of them.
Source: qiaojin/PubMedQA
Code and converter: Layered-Labs/benchbase
Items: 1,000 (train: 1,000)
Rejected source rows: 0 (see rejected.json)
Schema
Field
Type
Description
dataset_key
str
Benchmark identifier (pubmedqa)
hash
str
SHA256(question + answer text), stable across… See the full description on the dataset page: https://huggingface.co/datasets/Layered-Labs/benchbase-pubmedqa.benchbase-mmlu-medical
BenchBase MMLU Medical
MMLU Medical normalized into the BenchBase unified schema. Every BenchBase dataset shares the same
columns, so one loader and one evaluation loop cover all of them.
Source: cais/mmlu
Code and converter: Layered-Labs/benchbase
Items: 1,242 (test: 1,089, validation: 123, dev: 30)
Rejected source rows: 0 (see rejected.json)
Schema
Field
Type
Description
dataset_key
str
Benchmark identifier (mmlu_medical)
hash
str
SHA256(question… See the full description on the dataset page: https://huggingface.co/datasets/Layered-Labs/benchbase-mmlu-medical.pokedao-mew1a-training-data-layered
PokeDAO Mew-1A Training Dataset (Layered Architecture)
🧬 Project Mew-1A: The world's first AI training dataset specifically for Pokemon TCG pricing analysis, extracted from a production-ready layered database architecture.
Dataset Description
This dataset contains 10,000 high-quality training examples extracted from PokeDAO's layered database containing 116,744 market listings across multiple marketplaces.
Layered Architecture
The data is sourced from a… See the full description on the dataset page: https://huggingface.co/datasets/ChicoPanama/pokedao-mew1a-training-data-layered.benchbase-triage
BenchBase TRIAGE
TRIAGE normalized into the BenchBase unified schema. Every BenchBase dataset shares the same
columns, so one loader and one evaluation loop cover all of them.
Source: NLie2/TRIAGE
Code and converter: Layered-Labs/benchbase-med
Items: 86 (train: 86)
Rejected source rows: 1 (see rejected.json)
Schema
Field
Type
Description
dataset_key
str
Benchmark identifier (triage)
hash
str
SHA256(question + answer text), stable across option order… See the full description on the dataset page: https://huggingface.co/datasets/Layered-Labs/benchbase-triage.Qwen-Image-Layered-210-samples-redoqwen-layered-sessionsQwen-Image-Layered-210-samplesOmniPSD_Layered_PosterMAUVIKAT_QW_2512_Product_Posters_Qwen_Image_layered_overlayedMAUVIKAT_Product_Posters_Qwen_Image_layered_overlayed
MAUVIKAT (喵维卡特)_AD_Posters