datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
wds_objectnethot3d
HOT3D-Clips
This Hugging Face repository hosts HOT3D-Clips, a set of curated sub-sequences of the HOT3D dataset.
Download instructions for HOT3D-Clips and the full HOT3D dataset can be found here.
See HOT3D Toolkit for documentation of the data format and for Python utilities (for loading, undistorting fisheye images, rendering using fisheye cameras, etc.).
More details can be found in the HOT3D paper and BOP 2024 report.
RoadmapBench
RoadmapBench
A benchmark for evaluating AI coding agents on multi-target, long-horizon software development tasks derived from open-source project version upgrades.
Overview
RoadmapBench contains 115 tasks spanning 17 open-source repositories across 5 programming languages (Python, TypeScript, Go, Rust, C++). Each task requires an agent to implement multiple interdependent features that correspond to a real version upgrade of the target project.
Task Structure… See the full description on the dataset page: https://huggingface.co/datasets/benchmark-anon-2026/RoadmapBench.wds_imagenet_sketchBLINK
BLINK: Multimodal Large Language Models Can See but Not Perceive
🌐 Homepage | 💻 Code | 📖 Paper | 📖 arXiv | 🔗 Eval AI
This page contains the benchmark dataset for the paper "BLINK: Multimodal Large Language Models Can See but Not Perceive"
Introduction
We introduce BLINK, a new benchmark for multimodal language models (LLMs) that focuses on core visual perception abilities not found in other evaluations. Most of the BLINK tasks can be solved by humans “within a… See the full description on the dataset page: https://huggingface.co/datasets/BLINK-Benchmark/BLINK.MultiBanana-Benchmark🍌 MultiBanana: A Challenging Benchmark for Multi-Reference Text-to-Image Generation 🍌
CVPR 2026 (Main)
This repository provides the datasets for
“MultiBanana: A Challenging Benchmark for Multi-Reference Text-to-Image Generation” by Yuta Oshima, Daiki Miyake, Kohsei Matsutani, Yusuke Iwasawa, Masahiro Suzuki, Yutaka Matsuo and Hiroki Furuta
Paper Link
https://arxiv.org/abs/2511.22989
Github Repository
For the usage of this benchmark, please see Github… See the full description on the dataset page: https://huggingface.co/datasets/kohsei/MultiBanana-Benchmark.wds_imagenet-rGAIA
GAIA dataset
GAIA is a benchmark which aims at evaluating next-generation LLMs (LLMs with augmented capabilities due to added tooling, efficient prompting, access to search, etc).
We added gating to prevent bots from scraping the dataset. Please do not reshare the validation or test set in a crawlable format.
Data and leaderboard
GAIA is made of more than 450 non-trivial question with an unambiguous answer, requiring different levels of tooling and autonomy to… See the full description on the dataset page: https://huggingface.co/datasets/gaia-benchmark/GAIA.wds_imagenet-abenchmark
SLM Lab
Modular Deep Reinforcement Learning framework in PyTorch.
Companion library of the book Foundations of Deep Reinforcement Learning.
Documentation · Benchmark Results
NOTE: v5.0 updates to Gymnasium, uv tooling, and modern dependencies with ARM support - see CHANGELOG.md.
Book readers: git checkout v4.1.1 for Foundations of Deep Reinforcement Learning code.
BeamRider
Breakout
KungFuMaster
MsPacman
Pong
Qbert
Seaquest
Sp.Invaders… See the full description on the dataset page: https://huggingface.co/datasets/SLM-Lab/benchmark.wds_imagenet1ksvg-benchmark
Rapidata Static SVG Generation Benchmark
Built by Rapidata.
This dataset contains 1,918,367 human responses, collected with the
Rapidata Python SDK, comparing how well 42 frontier LLMs generate
static SVGs from text prompts. Each row is a head-to-head comparison between two models' renders of
the same prompt, scored by human annotators on one of three questions (Preference, Coherence, Alignment).
The SVGs are produced as raw <svg> markup by the models, rasterized to 768×768 PNGs… See the full description on the dataset page: https://huggingface.co/datasets/Rapidata/svg-benchmark.PDE_Inverse_Problem_Benchmarking
PDEInvBench: A Comprehensive Dataset and Design Space Exploration of Neural Networks for PDE Inverse Problems
This is the official dataset for the paper PDEInvBench: A Comprehensive Dataset and Design Space Exploration of Neural Networks for PDE Inverse Problems.
Code: GitHub - ASK-Berkeley/PDEInvBench
Sample Usage
You can use the provided script from the codebase to batch download the data:
pip install huggingface_hub
python3 huggingface_pdeinv_download.py --dataset… See the full description on the dataset page: https://huggingface.co/datasets/DabbyOWL/PDE_Inverse_Problem_Benchmarking.wds_imagenetv2oxford_flowers102GOAI-2026wds_fer2013MOVA_benchmark_for_arena
MOVA Benchmark for Arena
This is the benchmark used for the subjective arena experiments of MOVA (MOVA: Towards Scalable and Synchronized Video–Audio Generation). All prompts are rewritten by the workflow introduced in the paper.
Paper: MOVA: Towards Scalable and Synchronized Video–Audio Generation
Code: https://github.com/OpenMOVA/MOVA
Overview
The benchmark contains 732 samples in total, organized into two subsets:
Subset
Samples
MOVA-Bench
132… See the full description on the dataset page: https://huggingface.co/datasets/zhiyuzhang-0212/MOVA_benchmark_for_arena.Food_Portion_Benchmark
Food Portion Benchmark (FPB) Dataset
The Food Portion Benchmark (FPB) is a comprehensive dataset and benchmark suite for multi-task food scene understanding, combining food detection and portion size (weight) estimation. It was introduced to support research in dietary analysis, nutrition tracking, and food computing. The dataset is built with high-quality annotations and evaluated using an extended YOLOv12-based multi-task model .
📦 Dataset Overview
Total images:… See the full description on the dataset page: https://huggingface.co/datasets/issai/Food_Portion_Benchmark.navverse-benchmark
NavVerse Benchmark
This repository hosts the NavVerse dataset.
Runnable scene release
navverse_v1_july23.tar.gz is the current runnable scene package. It contains:
52 VC+ outdoor scenes and 52 connected indoor/outdoor scenes under vc_plus/;
30 GRScenes commercial navigation scenes under grscenes_commercial/;
the corresponding prebuilt navmesh/ assets;
a relative nvidia -> vc_plus/nvidia link for the bundled CloudySky runtime lighting.
From a NavVerse-Benchmark… See the full description on the dataset page: https://huggingface.co/datasets/tccoin/navverse-benchmark.Awesome_Spatial_VQA_BenchmarksIDEAL-Scenes
IDEAL-Bench: Indoor Dataset for Evaluating Analysis by 3D Layout Reasoning
IDEAL-Bench is an evaluation suite that requires VLMs to predict structured 3D layouts on photorealistic indoor scenes across 10 room types, scored along five numerical dimensions (scene validity, physical plausibility, geometric accuracy, object recognition, and grid layout) and a perceptual render-and-compare protocol.
Built on IDEAL-Scenes - 1,000 procedurally generated, re-renderable Blender scenes… See the full description on the dataset page: https://huggingface.co/datasets/IDEAL-Benchmark/IDEAL-Scenes.MMIU-Benchmark
Dataset Card for MMIU
Repository: https://github.com/OpenGVLab/MMIU
Paper: https://arxiv.org/abs/2408.02718
Project Page: https://mmiu-bench.github.io/
Point of Contact: Fanqing Meng
Introduction
MMIU encompasses 7 types of multi-image relationships, 52 tasks, 77K images, and 11K meticulously curated multiple-choice questions, making it the most extensive benchmark of its kind. Our evaluation of 24 popular MLLMs, including both open-source and proprietary models… See the full description on the dataset page: https://huggingface.co/datasets/FanqingM/MMIU-Benchmark.FedRemoteSensing_Benchmark
Dataset README
1. General Information
Number of Labels: There are a total of 5 labels, namely: Agriculture, Bareland, Forest, Residential, and River.
Number of Clients: The dataset consists of 100 clients.
Data Volume per Client: Each client contains approximately 350 tif format images.
2. Data Sources
All the images are collected from 6 different datasets, which are as follows:
Eurosat
UC Merced Land Use Dataset
AID
NWPU - RESISC45
WHU-RS19
NaSC-tg2
The data… See the full description on the dataset page: https://huggingface.co/datasets/billhdzhao/FedRemoteSensing_Benchmark.figma-slide-benchmark
Figma Slide Editing Benchmark
Benchmark accompanying our EMNLP 2026 Industry Track (Main) accepted paper "ACE: A
Self-Correcting Agentic Canvas Editor for Multi-Slide Presentation
Automation".
📄 Paper: https://arxiv.org/pdf/2608.24103
💻 Code: https://github.com/BloomBerry/agentic-canvas-editor
Overview
Each benchmark item is a slide-editing task defined as a pair of Figma Slides
documents:
*_TestA — the input deck the agent starts from.
*_GroundTruthA — the… See the full description on the dataset page: https://huggingface.co/datasets/BloomBerry/figma-slide-benchmark.NL3D-Synth-Benchmark
Dataset Card for NL3D-Synth-Benchmark
Dataset Summary
NL3D-Synth-Benchmark is the synthetic evaluation split of NL3D, a synthetic-real dataset and benchmark for non-Lambertian 3D reconstruction. This split is designed to provide a compact, controlled, and reproducible benchmark for evaluating 3D perception methods on challenging non-Lambertian materials.
The benchmark contains 16 synthetic scenes rendered with physically based material models and full 3D background… See the full description on the dataset page: https://huggingface.co/datasets/NL3D/NL3D-Synth-Benchmark.UAVDT-Benchmark-Minvoice-extraction-benchmark
Invoice Extraction Benchmark v1
A synthetic test set for invoice data extraction (invoice OCR, intelligent document
processing, accounts-payable capture): 181 documents with answer keys and a scorer.
Run any invoice reader over the documents, write its output as one JSON file, and score it
field by field.
Source, scorer and generator: https://github.com/DrewKraken/invoice-extraction-benchmark
(this dataset is a mirror of corpus/v1/ there; the GitHub repository is canonical).
Who… See the full description on the dataset page: https://huggingface.co/datasets/drew-ipp/invoice-extraction-benchmark.synthetic-medical-document-recognition-benchmark
Synthetic Medical Document Recognition Benchmark
This dataset contains synthetic, English-language medical records rendered as
documents for evaluating automated data extraction and de-identification
systems. Each synthetic patient has a longitudinal FHIR R4 record and multiple
visual representations derived from that record.
Every rendered document is clearly marked as synthetic. This makes the dataset
suitable for manual testing, product demonstrations, and workflows that… See the full description on the dataset page: https://huggingface.co/datasets/morzel85/synthetic-medical-document-recognition-benchmark.mtg-urna-benchmark
38,627 Magic: The Gathering cards, one per oracle id, the scan and the rules text of each, packed into single .urna files that answer text and image queries from memory-mapped bytes: no server, no Python at read time. Ten such files live here. They carry the same cards, the same text and the same content hash; what differs is how the 4 GB of JPEG was encoded inside, and which image models embedded it.
The file to start with is release/v0.3/stills-5models. It is the only one with the models… See the full description on the dataset page: https://huggingface.co/datasets/brennercruvinel/mtg-urna-benchmark.