factorpol.alphas_training

This module provides functionalities to derive and obtain typed polarizabilities from QM reference data.

This module uses Ray to distribute the optimization process. The optimization process is performed by scipy.optimize.minimize with Nelder-Mead method.

Functions

optimize_alphas(worker_list[, solved, num_cpus])	A function to optimize the polarizability of a dataset to reference QM ESPs Atomic units
optimize_alphas_fast(worker_list[, solved, ...])	** This method is extremely experimental and not recommended for production use **

Classes

AlphaData(database_name, dataset, ...[, ...])	A class to prepare reference QM ESPs for optimization of polarizability
AlphaWorker(record, off_forcefield, ...[, ...])	A class used to derive polarizability from QM ESPs.
AlphasTrainer(workers, prior[, ...])	Top level optimizer to train polarizability parameters