Datasets:

bgce
/

cooldata-v2

Size:

10K<n<100K

DOI:

License:

Dataset card Files Files and versions

xet

Community

cooldata-v2 / README.md

peteole

Update README.md

5298a3c verified 9 months ago

preview code

raw

history blame contribute delete

1.69 kB

metadata

pretty_name: CoolData - A 3d Industrial-Grade Flow Field Dataset for Electronics Cooling
size_categories:
  - 10K<n<100K
license: cc-by-nc-4.0

Cooldata - A Large-Scale Electronics Cooling 3D Flow Field Dataset

Cooldata is a large-scale electronics cooling dataset, containing over 60k stationary 3D flow fields for a diverse set of geometries, simulated with the commercial solver Simcenter STAR-CCM+. This library can be used to acccess the dataset and streamline its application in machine learning tasks.

Find the documentation at cooldata.readthedocs.io.

Features

Data Storage: Organized in folders containing .cgns files for compatibility with computational fluid dynamics tools.
PyVista Integration: Access to dataset samples as PyVista objects for easy 3D visualization and manipulation.
Graph Neural Network Support:
- DGL Support:
  - Surface and volume data in mesh format.
  - 3D visualization of samples and predictions.
  - L2 loss computation and aggregate force evaluation for model training.
- PyG Support: Implementing functionalities similar to DGL.
Hugging Face Integration: Direct dataset loading from Hugging Face.
Voxelized Flow Field Support: Facilitates image processing-based ML approaches.
Comprehensive Metadata Accessibility: All metadata is accessible through the library.

Installation

Run

pip install cooldata

If you want to use the DGL support, you also need to install the DGL library, as documented here.