# The Well: Turbulence with Gravity and Radiative Cooling in the Interstellar Medium

## Facts

- Provider: Polymathic AI (The Well); Keiya Hirashima, University of Tokyo and Flatiron Institute CCA
- License: CC-BY-4.0 (https://arxiv.org/html/2412.00568)
- Size: 90 files
- Format: HDF5 (in sampled files)
- Stored at: Hugging Face, https://huggingface.co/datasets/polymathic-ai/turbulence_gravity_cooling
- Access: public, readable without an account
- Category: Physics
- Source: https://arxiv.org/abs/2412.00568
- Page: https://scigantic.com/archive/bc47f4d1-24b2-5a50-98f1-de02ca8f1b6c
- Keywords: Physics, HDF5, Hugging Face

## About

3D simulations of a turbulent, self-gravitating fluid modelling the interstellar medium (ISM) of galaxies at different evolutionary stages, where dense filaments form that go on to make stars. The hydrodynamics with gravity and radiative heating and cooling is solved with ASURA-FDPS smoothed particle hydrodynamics (SPH) and gridded to 64 by 64 by 64 cubes over 50 time steps, spanning one free-fall time. Fields are pressure, density, temperature and velocity. The datasheet lists 2,700 trajectories: 27 parameter sets (initial temperature 10, 100 or 1000 K; hydrogen number density 44.5, 4.45 or 0.445 per cc; metallicity Z0, 0.1 Z0 or 0) times 100 runs, with open boundaries. Listing data/ (measured 2026-11-13) shows train/ (27 files, 680 GB, one 25.2 GB file per parameter set), valid/ and test/ (27 files, 85.8 GB each); about 852 GB in the repository. Files are HDF5 in the common The Well format; open them with the_well package (WellDataModule with base path hf://datasets/polymathic-ai/) or directly with h5py. Part of The Well (Ohana et al., NeurIPS 2024 Datasets and Benchmarks, arXiv:2412.00568); released under CC-BY-4.0, cite The Well and the simulation's source papers.

## Load the data

```python
from huggingface_hub import hf_hub_download
path = hf_hub_download(repo_id="polymathic-ai/turbulence_gravity_cooling", filename="data/test/turbulence_gravity_cooling_rho0_0.445_Z_0.1_T0_10.hdf5", repo_type="dataset")
print(path)
```

## Data structure

File types in sampled files: HDF5

Example file: `data/test/turbulence_gravity_cooling_rho0_0.445_Z_0.1_T0_10.hdf5`

## Open it

- In a hosted notebook with the data mounted: https://scigantic.com/archive/bc47f4d1-24b2-5a50-98f1-de02ca8f1b6c
- Full record as JSON (schema card included): https://api.scigantic.com/api/archives/bc47f4d1-24b2-5a50-98f1-de02ca8f1b6c
- Over MCP: `get_schema_card` with id `bc47f4d1-24b2-5a50-98f1-de02ca8f1b6c` at https://api.scigantic.com/mcp

## Related topics

- [HDF5 datasets](https://scigantic.com/topics/hdf5)
- [Physics datasets](https://scigantic.com/topics/physics)
