Datasets

Note: placeholder links and text. The final archives have not been generated yet. Give it a couple months

Intended use and Licensing

The datasets are distributed solely for non-commercial academic research. Permitted uses are limited to direct academic research activities conducted without commercial intent.
Use for training, fine-tuning, or otherwise improving large language models or other AI systems is expressly forbidden.
These limitations apply on top of the license specified in the footer of this page.
If unsure, contact information is available on the Home page.

Microstructures

These datasets all contain the same type of structures. All structures have the following properties:

Download

Using OTBV encoding:Raw volume data in a HDF5 format. Default zip compression.

Only the 1k dataset has been manually checked for quality due to size.
All data is synthetic.
The generation algorithm is open-source [link after published].