# The Well: Turbulent Radiative Mixing Layer (3D)

## Facts

- Provider: Polymathic AI (The Well); Drummond Fielding, Flatiron Institute CCA and Cornell University
- License: CC-BY-4.0 (https://arxiv.org/html/2412.00568)
- Size: 37 files
- Format: HDF5 (in sampled files)
- Stored at: Hugging Face, https://huggingface.co/datasets/polymathic-ai/turbulent_radiative_layer_3D
- Access: public, readable without an account
- Category: Physics
- Source: https://arxiv.org/abs/2412.00568
- Page: https://scigantic.com/archive/d992375b-d206-5928-990c-80ab5b6a6930
- Keywords: Physics, HDF5, Hugging Face

## About

The 3D counterpart of the 2D radiative mixing layer set: cold dense gas below hot dilute gas in relative subsonic motion, seeded with small-scale noise and unstable to the Kelvin-Helmholtz instability, so that mixed intermediate-temperature gas cools rapidly. Computed with Athena++ on 256 by 128 by 128 cubes (x and y in [-0.5, 0.5], z in [-1, 1]), periodic in x and y and zero-gradient in z; each trajectory has 101 time steps. Fields are density, pressure and velocity. The datasheet lists 90 trajectories: 10 seeds for each of 9 cooling times t_cool from 0.03 to 3.16. Listing data/ (measured 2026-11-13) shows train/ (9 files, 610 GB, 67.8 GB per cooling time), valid/ and test/ (9 files, 76.3 GB each); about 763 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/turbulent_radiative_layer_3D", filename="data/test/turbulent_radiative_layer_tcool_0.03.hdf5", repo_type="dataset")
print(path)
```

## Data structure

File types in sampled files: HDF5

Example file: `data/test/turbulent_radiative_layer_tcool_0.03.hdf5`

## Open it

- In a hosted notebook with the data mounted: https://scigantic.com/archive/d992375b-d206-5928-990c-80ab5b6a6930
- Full record as JSON (schema card included): https://api.scigantic.com/api/archives/d992375b-d206-5928-990c-80ab5b6a6930
- Over MCP: `get_schema_card` with id `d992375b-d206-5928-990c-80ab5b6a6930` at https://api.scigantic.com/mcp

## Related topics

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