A python module centered around developing, testing, and deploying machine-learning algorithms to quantum chemical datasets. We have a focus on wave function-based descriptors, with capabilities for building representations from ab initio electronic structure data such as wave function amplitudes or reduced density matrices.
Beyond representation generation, some scikit-learn functions are wrapped for ease-of-use with the Dataset class, such as k-means clustering and kernel ridge regression. Defaults are chosen based on performance; however, additional options are available, and scikit-learn may also be used directly. See docstrings for details.
To install, run:
pip install -e .To test, run:
py.testfrom the base mlqm/ directory.