datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
fluidgym-datamusan
MUSAN: A Music, Speech, and Noise Corpus
MUSAN is a corpus of music, speech, and noise recordings designed for training models for voice activity detection and music/speech discrimination. This is a comprehensive collection suitable for various audio processing tasks.
Dataset Structure
The dataset is organized into three main categories:
1. Music (~42 hours)
Subcategories: Classical, Pop/Rock, Jazz, and more
Sources: Free Music Archive, Jamendo, and others… See the full description on the dataset page: https://huggingface.co/datasets/FluidInference/musan.fleurs-full
FLEURS Full - Test Set for ASR Benchmarking
Complete test set of Google FLEURS for all 30 languages supported by Qwen3-ASR, prepared for benchmarking with FluidAudio.
Languages (30)
Asian Languages (13)
Code
Language
Samples
cmn_hans_cn
Chinese (Mandarin)
945
yue_hant_hk
Cantonese
819
ja_jp
Japanese
650
ko_kr
Korean
382
vi_vn
Vietnamese
857
th_th
Thai
1,021
id_id
Indonesian
687
ms_my
Malay
749
hi_in
Hindi
418
ar_eg
Arabic (Egyptian)… See the full description on the dataset page: https://huggingface.co/datasets/FluidInference/fleurs-full.ami-corpus-mirror
AMI Corpus Mirror
Mirror of the subset of the AMI Meeting Corpus used by
FluidAudio diarization benchmarks. Hosted here so CI and
local benchmark runs do not depend on the availability of the upstream groups.inf.ed.ac.uk server
(see FluidAudio#752).
Contents
annotations/ami_public_manual_1.6.2.zip — AMI public manual annotations v1.6.2
(repackaged from the official archive; identical content, including segments/, words/,
corpusResources/meetings.xml)… See the full description on the dataset page: https://huggingface.co/datasets/FluidInference/ami-corpus-mirror.THCHS-30-tests
THCHS-30 Test Set
THCHS-30 test split for Mandarin Chinese speech recognition benchmarking.
Dataset Info
Language: Mandarin Chinese (zh-CN)
Samples: 2,495
Speakers: 10
Sample Rate: 16 kHz
License: Apache 2.0
Usage
from datasets import load_dataset
# After uploading to HuggingFace
dataset = load_dataset("your-username/thchs30-test")
# Example
print(dataset['train'][0])
# {
# 'audio': {'array': [...], 'sampling_rate': 16000, 'path': 'audio/D11_750.wav'},
#… See the full description on the dataset page: https://huggingface.co/datasets/FluidInference/THCHS-30-tests.HUGGER-Unified-Gravity-Fluid-Framework
🌍 H.U.G.G.E.R: Heuristic Universal Grid & Gravity Equilibrium Rendering Tensor
This repository serves as an open academic archive and tensor-specification benchmark for generalized tensor standards, designed to resolve non-linear computational collapse and topological pole singularities in high-performance CFD and planetary atmospheric models.
It acts as the Macroscopic Gravitational Backbone, perfectly entangled with the microscopic Topological Zero Tensor (TZT)… See the full description on the dataset page: https://huggingface.co/datasets/jskresearch/HUGGER-Unified-Gravity-Fluid-Framework.FluidNexusDatasets
FluidNexus: 3D Fluid Reconstruction and Prediction From a Single Video
Yue Gao*, Hong-Xing "Koven" Yu*, Bo Zhu, Jiajun Wu
Stanford University; Microsoft; Georgia Institute of Technology
* denotes equal contribution
FluidNexus-Smoke and FluidNexus-Ball
Our FluidNexus-Smoke and FluidNexus-Ball datasets each include 120 scenes. Every scene contains 5 synchronized multi-view videos, with cameras arranged along a horizontal arc of approximately 120°.
FluidNexusSmoke… See the full description on the dataset page: https://huggingface.co/datasets/terrant20/FluidNexusDatasets.samples
Fluidverse Samples
This dataset provides sample files for all of the datasets published by the Fluidverse Organization.
For each dataset, two trajectories are provided - one with a single bubble / droplet and one multi-droplet / bubble trajectory with five bubbles / droplets.
Note: This is not a standalone dataset with a test and train split. This huggingface dataset serves solely as a preview for the complete datasets listed below.
Dataset samples
Dataset… See the full description on the dataset page: https://huggingface.co/datasets/FluidVerse/samples.OpenSeeSimE-Fluid
OpenSeeSimE-Fluid: Engineering Simulation Visual Question Answering Benchmark
Dataset Summary
OpenSeeSimE-Fluid is a large-scale benchmark dataset for evaluating vision-language models on computational fluid dynamics (CFD) simulation interpretation tasks. It contains approximately 98,000 question-answer pairs across parametrically-varied fluid simulations including turbulent flow, heat transfer, and complex flow patterns.
Purpose
While vision-language models… See the full description on the dataset page: https://huggingface.co/datasets/cmudrc/OpenSeeSimE-Fluid.JSUT-basic5000
JSUT (Japanese Speech Corpus) - Test Subset
A test subset of the JSUT corpus containing 500 Japanese utterances from the basic5000 dataset (BASIC5000_4501-5000).
Dataset Structure
jsut_ver1.1/
└── basic5000/
├── wav/ # WAV audio files (500 files, 48kHz)
├── transcript_utf8.txt # Transcriptions
└── recording_info.txt # Recording dates
File Formats
transcript_utf8.txt
BASIC5000_4501:だが、エーアイセンター稼動を快く思わない...… See the full description on the dataset page: https://huggingface.co/datasets/FluidInference/JSUT-basic5000.fluidgym-experiments
FluidGym Experiments
Paper | GitHub | Documentation
FluidGym is a standalone, fully differentiable benchmark suite for reinforcement learning (RL) in active flow control (AFC). Built entirely in PyTorch on top of the GPU-accelerated PICT solver, it provides standardized evaluation protocols and diverse environments for systematic comparison of control methods.
This repository contains the training and test datasets with results for all experimental runs presented in the paper.… See the full description on the dataset page: https://huggingface.co/datasets/safe-autonomous-systems/fluidgym-experiments.librispeechsample-page-assets
Sample-page assets
Files the cards of the Fluid Concepts sample datasets on the Hub (ToolTalk, Multimodal Expert Instruction, Multimodal Peer Collaboration)
need to show to visitors who have not requested access yet, which the gated sample repositories cannot serve themselves:
the card banners (*.png);
a public copy of each sample repository's TECHNICAL.md and TERMS.md, under the repository's name, refreshed on every push of that repository.
Nothing else is published here. The… See the full description on the dataset page: https://huggingface.co/datasets/fluid-concepts/sample-page-assets.OpenSeeSimE-Fluid-Small
OpenSeeSimE-Fluid-Small
A stratified 10% subset of cmudrc/OpenSeeSimE-Fluid for evaluating vision-language models at a reduced compute footprint while preserving the joint distribution of simulation type, question type, media type, and question id.
Subset Provenance
Parent dataset: cmudrc/OpenSeeSimE-Fluid (98,326 rows total)
Rows in this subset: 9,881 (10.05% of parent)
Source classes: Bent Pipe, Converging Nozzle, Heat Exchanger, Heat Sink, Mixing Pipe
Parquet shards:… See the full description on the dataset page: https://huggingface.co/datasets/cmudrc/OpenSeeSimE-Fluid-Small.fleurs
FLEURS Test Dataset
Reorganized FLEURS test dataset with audio and transcripts together.
Structure
fleurs-test/
├── en_us/
│ ├── en_us_0000.wav
│ ├── en_us_0001.wav
│ ├── ...
│ ├── en_us.trans.txt (LibriSpeech format)
│ ├── en_us.csv (detailed metadata)
│ └── en_us.json (JSON metadata)
├── fr_fr/
│ └── ...
└── ...
Languages
bg_bg: 350 test samples
cs_cz: 350 test samples
da_dk: 930 test samples
de_de: 350 test samples
el_gr: 650… See the full description on the dataset page: https://huggingface.co/datasets/FluidInference/fleurs.deltav-fluidtooltalk-samples
ToolTalk Samples - High Quality Duplex Speech and Tool-calling in Customer Service Domain
Two people improvise realistic customer-service calls while one operates a live, stateful tool environment—with synchronized speaker-separated audio, tool calls, and outcomes.
▶ Listen to Clean · ▶ Listen to Noisy · Discuss the full dataset
In this sample: 26 calls · 90.7 minutes · 7 sample domains · 207 tool calls
Technical specs: 48 kHz / 32-bit PCM speaker-separated source… See the full description on the dataset page: https://huggingface.co/datasets/fluid-concepts/tooltalk-samples.FluidNexusDatasets
FluidNexus: 3D Fluid Reconstruction and Prediction From a Single Video
Yue Gao*, Hong-Xing "Koven" Yu*, Bo Zhu, Jiajun Wu
Stanford University; Microsoft; Georgia Institute of Technology
* denotes equal contribution
FluidNexus-Smoke and FluidNexus-Ball
Our FluidNexus-Smoke and FluidNexus-Ball datasets each include 120 scenes. Every scene contains 5 synchronized multi-view videos, with cameras arranged along a horizontal arc of approximately 120°.
FluidNexusSmoke… See the full description on the dataset page: https://huggingface.co/datasets/yuegao/FluidNexusDatasets.cv-corpus-25.0-ja
Mozilla Common Voice 25.0 - Japanese Test Set (Complete)
Dataset Description
Complete Japanese test set from Mozilla Common Voice Corpus 25.0. This dataset contains all 9,019 validated test samples, compared to the partial 2,334-sample version previously available on HuggingFace.
Key Features
Size: 9,019 validated test utterances
Coverage: 100% of official Common Voice 25.0 Japanese test split
Multi-speaker: Diverse set of speakers with demographic metadata… See the full description on the dataset page: https://huggingface.co/datasets/FluidInference/cv-corpus-25.0-ja.fluid-benchmarking
Fluid Language Model Benchmarking
This dataset provides IRT models for ARC Challenge,
GSM8K,
HellaSwag,
MMLU,
TruthfulQA, and
WinoGrande.
Furthermore, it contains
results for pretraining checkpoints of Amber-6.7B,
K2-65B,
OLMo1-7B,
OLMo2-7B,
Pythia-2.8B, and
Pythia-6.9B, evaluated on these six benchmarks.
🚀 Usage
For utilities to use the dataset and to replicate the results from the paper, please see the corresponding GitHub… See the full description on the dataset page: https://huggingface.co/datasets/allenai/fluid-benchmarking.friend-bench
Can a model — or a human — tell how two people are related from a 20-second clip of how they interact?
🌐 Built on Seamless Interaction
FriendBench is a suite of benchmarks for social perception from thin-slice dyadic
interaction — inferring facts about two people's relationship from a brief clip of how they
interact, built on the Seamless Interaction
dataset. Each released set is a config of this repository.
🎧 Multi-modal — text, audio, and video for every clip
🎯 Objective label —… See the full description on the dataset page: https://huggingface.co/datasets/fluid-concepts/friend-bench.Reasoning_as_Fluid
Reasoning as Fluid: The Minimal Primitives and Inevitable Collapse of Linear Space Representation
Author: Zixi "Oz" Li (Independent Researcher)
Publication Date: November 2025
DOI: 10.57967/hf/7081
arXiv: (Submitted)
PDF: fluid_reasoning.pdf
📋 Abstract
We establish that reasoning is fundamentally a fluid-like dynamical process on constrained manifolds, not a computation in linear vector spaces. Through rigorous mathematical proof, we demonstrate:
Minimal Reasoning… See the full description on the dataset page: https://huggingface.co/datasets/OzTianlu/Reasoning_as_Fluid.multimodal-peer-collaboration-samples
Multimodal Peer Collaboration Samples - Embodied Map Task with Two Camera Angles
Two non-experts collaborate to build working circuits under asymmetric information: the instructor has the manual, the student has the components, and synchronized audio and dual-camera video capture how shared understanding emerges.
▶ Watch the interactions · See Expert Instruction samples · Discuss the full collection
Sister collection: Expert Instruction, a teacher and a student in… See the full description on the dataset page: https://huggingface.co/datasets/fluid-concepts/multimodal-peer-collaboration-samples.2D_SABW_OOOO
2D Shock-Induced Air Bubble Collapse in Water with Open BC
Description
The dataset captures the time-evolving behavior of 2D cylindrical air bubbles subjected to an external shock wave in water.
The interaction with the shock wave results in a collapse of the air bubble.
Here we investigate a scenario with open boundary conditions at every wall.
About the data
Metadata
Description
Solver
ALPACA
PDE
2D compressible Euler equations… See the full description on the dataset page: https://huggingface.co/datasets/FluidVerse/2D_SABW_OOOO.multimodal-expert-instruction-samples
Multimodal Expert Instruction Samples - Musical Instrument Lessons with Channel-separated Audio and Video
A music teacher and a student work through two one-on-one lessons: both voices and both instruments on separate tracks, the student on camera, with the lesson plans, the instructions given to each side and both sides' post-lesson ratings alongside.
▶ Watch the lessons · See Peer Collaboration samples · Discuss the full collection
Sister collection: Peer Collaboration… See the full description on the dataset page: https://huggingface.co/datasets/fluid-concepts/multimodal-expert-instruction-samples.2D_SDBA_SSOO
2D Shock-Induced Droplet Breakup in Air with Symmetric BC
Description
The dataset captures the time-evolving behavior of 2D cylindrical droplets subjected to an external shock wave in air.
The interaction with the shock wave results in two different breakup modes namely SIE and RTP depending on the weber number.
Here we investigate a scenario with symmetric boundary conditions at the north and south walls.
About the data
Metadata
Description… See the full description on the dataset page: https://huggingface.co/datasets/FluidVerse/2D_SDBA_SSOO.2D_SRBA_OOOO
2D Shock-Induced R22 Bubble Collapse in Air with Open BC
Description
The dataset captures the time-evolving behavior of 2D cylindrical R22 bubbles subjected to an external shock wave in air.
The interaction with the shock wave results in a collapse of the R22 bubble.
Here we investigate a scenario with open boundary conditions at every wall.
About the data
Metadata
Description
Solver
ALPACA
PDE
2D compressible Euler equations
Dimension… See the full description on the dataset page: https://huggingface.co/datasets/FluidVerse/2D_SRBA_OOOO.FluidSimulation
SURF: A Generalisation Benchmark for GNNs Predicting Fluid Dynamics
SURF, is a benchmark designed to test the generalization of learned graph-based fluid simulators. The benchmark consists of seven independent datasets:
Base
Turned
Topo
Range
Dynamic
Full
FullFiner
Each dataset is available as separate *.zip file and consists of at least 1200 2D incompressible fluid flow simulations with 300 timesteps.
The data structure is as follows:
folder: dataset_name
folders: dpx
files:… See the full description on the dataset page: https://huggingface.co/datasets/SURF-FluidSimulation/FluidSimulation.fluid-reasoning-representation-phase1
Fluid Reasoning Representation - Phase 1 Multi-Model + Cross-Domain Sweep
Phase 1 artifacts for the ARR 2026 rebuttal of Fluid Reasoning Representation
(Hook et al.). This dataset extends the original QwQ x Mystery Blocksworld
study with:
Second large reasoning model: Llama-3.3-Nemotron-Super-49B-v1
Two new domains: Mystery Logistics (PDDL Logistics with obfuscated
action / predicate vocabulary) and GSM8K-Renamed (math word problems with
surface noun + verb obfuscation).
C3 causal… See the full description on the dataset page: https://huggingface.co/datasets/AIM-Intelligence/fluid-reasoning-representation-phase1.OpenSeeSimE-Fluid-Mini
OpenSeeSimE-Fluid-Mini
A stratified 1% subset of cmudrc/OpenSeeSimE-Fluid for evaluating vision-language models at a reduced compute footprint while preserving the joint distribution of simulation type, question type, media type, and question id.
Subset Provenance
Parent dataset: cmudrc/OpenSeeSimE-Fluid (98,326 rows total)
Rows in this subset: 1,040 (1.06% of parent)
Source classes: Bent Pipe, Converging Nozzle, Heat Exchanger, Heat Sink, Mixing Pipe
Parquet shards: 3… See the full description on the dataset page: https://huggingface.co/datasets/cmudrc/OpenSeeSimE-Fluid-Mini.