datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
GMFlow_raw_simulation
GMFlow Raw Simulation
Overview
GMFlow Raw Simulation is a large-scale open dataset of high-fidelity three-dimensional earthquake ground-motion simulations generated for the GMFlow project. It contains both point-source and finite-rupture earthquake scenarios for San Francisco Bay Area, covering magnitudes Mw 4.4, Mw 6, and Mw 7.
The simulations are performed on the supercomputer Perlmutter at the National Energy Research Scientific Computing Center (NERSC), with a total… See the full description on the dataset page: https://huggingface.co/datasets/Yaozhong/GMFlow_raw_simulation.Paired_Compressible_Boussinesq_Flow_Simulation_with_Random_Temperature_BCs
Paired Compressible / Boussinesq Flow with Random Temperature BCs
📄 Paper: A Neural Surrogate Approach for Simulating Natural Convection
Problems (arXiv:2606.25259) — Nurshat Menglik,
Alex Shao, David Hyde.
10,000 matched pairs of 2D natural-convection simulations of the differentially
heated square cavity under randomized wall-temperature boundary conditions. Each
sample solves the same problem twice — once with the Boussinesq model and once
with the fully compressible model —… See the full description on the dataset page: https://huggingface.co/datasets/NurshatMenglik/Paired_Compressible_Boussinesq_Flow_Simulation_with_Random_Temperature_BCs.Simulation_asset_datasetsimulation-package
Simulation runtime: the third-party half of the AVSim data generation package
This repository holds the third-party runtime of the Simulation data generation package, version 3: the pieces the pipeline needs that were not written by the authors. It contains no code and no data of the authors. It is not usable on its own: the private half of the package, AVSim/simulation, downloads this repository into the same directory at a pinned revision with its fetch_runtime.sh script and… See the full description on the dataset page: https://huggingface.co/datasets/AVSim/simulation-package.Downstream_Physics_Simulation
GeoPT
Project Page | Paper | GitHub
This repository contains the physics simulation data for the paper GeoPT: Scaling Physics Simulation via Lifted Geometric Pre-Training.
GeoPT is a unified model pre-trained on large-scale geometric data for general physics simulation, unlocking a scalable path for neural simulation.
Overview
GeoPT is evaluated on the following five simulation tasks.
Dataset
Mesh Size
Variable
Training
Test
Total Size
Source
DrivAerML… See the full description on the dataset page: https://huggingface.co/datasets/GeoPT/Downstream_Physics_Simulation.SimulationMetadata
PercepTax — Simulation Metadata
Supporting metadata for the PercepTax benchmark.
This repository holds the object taxonomy, object lists / placement rules, and
per-scene 3D annotations used to render and annotate the simulated split of PercepTax.
It is a collection of JSON/CSV/PNG files (not a load_dataset-style tabular dataset).
Contents
taxonomy/ # the physical-property taxonomy
├── taxonomy.json # full taxonomy… See the full description on the dataset page: https://huggingface.co/datasets/TaxonomyProject/SimulationMetadata.tabletop-simulation-rlds
Talbetop Simulation Dataset - RLDS Format
Bimanual tabletop simulation dataset, collected on Tabletop-Sim.
In total, 20GB.
git clone https://huggingface.co/datasets/jellyho/tabletop-simulation-rlds
cd tabletop-simulation-rlds
git lfs pull
tabletop-simulation-hdf5rim2d-simulations
RIM2D East Africa Flood Simulations
Event-based 2D hydraulic inundation simulations for the GHACOF flood storyline
workshop (May 2026), produced under the ICPAC E4DRR / CRAF'D project.
Cases
Each <case>/ directory is self-contained and can be re-run with RIM2D:
<case>/
├── input/
│ ├── dem.nc Copernicus GLO-30 DEM → reference grid
│ ├── buildings.nc Overture Maps building fraction raster
│ ├── channel_mask.nc channel mask… See the full description on the dataset page: https://huggingface.co/datasets/E4DRR/rim2d-simulations.SimulationImageheat3d-thermal-simulation
Heat3D Thermal Simulation Dataset: Synthetic 3D Heat-Conduction Data for Operator Learning
This repository is the main dataset entry point for the Heat3D project. It is
intended to collect multiple synthetic 3D heat-conduction subsets for operator
learning research.
The root index currently highlights:
subsets/v0_unitcube_demo/: a prototype UnitCube dataset for checking data
loading, 3D graph construction, training, and evaluation workflows.… See the full description on the dataset page: https://huggingface.co/datasets/133754144X/heat3d-thermal-simulation.Downstream_Physics_Simulation
GeoPT
Project Page | Paper | GitHub
This repository contains the physics simulation data for the paper GeoPT: Scaling Physics Simulation via Lifted Geometric Pre-Training.
GeoPT is a unified model pre-trained on large-scale geometric data for general physics simulation, unlocking a scalable path for neural simulation.
Overview
GeoPT is evaluated on the following five simulation tasks.
Dataset
Mesh Size
Variable
Training
Test
Total Size
Source… See the full description on the dataset page: https://huggingface.co/datasets/introvoyz041/Downstream_Physics_Simulation.foundation-model-results-simulation-consortiarfi-simulationssimulations
Verasonics ultrasound data (zea format)
This dataset contains raw ultrasound data acquired with Verasonics systems,
converted to the zea HDF5 format.
wflow.jl-simulationsthat-one-google-math-datasetapolocheese for poor format, it's because I Don't Care (i'm tired and still working)
data from: https://github.com/google-deepmind/mathematics_dataset
from huggingface_hub import snapshot_download
from datasets import load_dataset
import os
def get_all_files(directory):
file_paths = []
for root, dirs, files in os.walk(directory):
for name in files:
full_path = os.path.join(root, name)
file_paths.append(os.path.abspath(full_path))
return file_paths… See the full description on the dataset page: https://huggingface.co/datasets/midwestern-simulation/that-one-google-math-dataset.agent-simulations
Agent Simulations
53,971 synthetic agent trajectories generated by simulations
across 34 agent types. The rows include successful and failed
trajectories for supervised fine-tuning, preference work, reinforcement learning, and
evaluation.
NOTE: This is generated test and training data, not curated ground truth. Review and
filter it for your application before training or evaluation.
Included agents
airline, amazon, bank, browser, calendar, chewy, clinic, coding… See the full description on the dataset page: https://huggingface.co/datasets/while-ai/agent-simulations.laser-arc_welding_simulation_dataset_tecplot_formatdeadline-render-simulation-20260605152225
Deadline render simulation 20260605152225
Generated by simulate-deadline-render-result.ts
This dataset mirrors public data-pack render outputs from Physicl.
manual_simulations_100stackexchange_flattened7.6M threads of posts + answers + comments from stackexchange (omitting stackoverflow).
with the Llama2 tokenizer (32k vocab) this should come out to ~7.94GT
fly-sud-simulation
FlyWire-informed odor-reward simulation: individual-behavior V4b
The full predeclared validation FAILED. This is synthetic simulation data,
not measured fly behavior or a quantitative reproduction of Kaun et al. (2011).
Detailed results ·
Code and protocols
Findings and limitations
256 independently seeded validation flies, four conditions (paired, unpaired,
untrained, retrieval-DAN-silenced), two delays (30 min, 24 h), 32 flies per cell:
8 reciprocal replicate… See the full description on the dataset page: https://huggingface.co/datasets/Histochemichael/fly-sud-simulation.batch_simulations_300vrh3-part-placing-simulation-RVT_TD_BUDynamics_Simulation_Dataset
Dynamics Simulation Dataset
Dataset Summary
This dataset contains dynamics simulation trajectories computed with the Natural Coordinate Method (NCM). It is organized by benchmark case. Each case provides a training split and a test split, and each split contains multiple simulation runs with different initial conditions.
The dataset is intended for learning and evaluating dynamics models, trajectory prediction methods, control-aware models, and surrogate models… See the full description on the dataset page: https://huggingface.co/datasets/PerseusLjf/Dynamics_Simulation_Dataset.so100_simulation_trashsimulation-assetsmanual_simulations_100fci-neuron-simulations
FCI neuron simulations
The simulated input-output data of the 24 detailed compartmental neuron models compared in
Dendritic morphology and synaptic nonlinearities enhance functional complexity in human cortical neurons
Ido Aizenbud, Daniela Yoeli, David Beniaguev, Christiaan P. J. de Kock, Michael London, Idan Segev.
PNAS 123(28), e2533168123 (2026).
Code, neuron models and instructions: https://github.com/ido4848/FCI. The trained networks
that the paper's Functional Complexity… See the full description on the dataset page: https://huggingface.co/datasets/i-do-ai/fci-neuron-simulations.