Materials Project Time Splits

Warning

As of June 2023, mp-time-split has migrated to matbench-genmetrics as a namespace package

Download Materials Project time-splits for generative modeling benchmarking via the mp_time_split Python package. Download and store a snapshot dataset of experimentally verified Materials Project entries sorted by earliest publication year of the associated literature references. Alternatively, fetch your own dataset directly from Materials Project using your own search criteria. The snapshot dataset, MPTS-52, contains only experimentally verified Materials Project entries with no more than 52 sites) and acts as an extension to the state-of-the-art materials generative model introduced by Xie et al. via the CDVAE package. We recommend that in addition to the three CDVAE benchmark datasets, you also use the MPTS-52 dataset and corresponding cross-validation and final test splits for model comparison and benchmarking. MPTS-52 can be used with the metrics introduced in CDVAE’s compute_metrics.py script (see CDVAE instructions). Check out our quick start page for more details on how to use mp_time_split. Enjoy!

See also: matbench-genmetrics, a materials benchmarking platform for generative models and xtal2png, a tool converting between a crystal structure and a PNG image representation for generative modeling

Star Follow @sgbaird Issue Discuss

Contents

Indices and tables