gninatorch.training module

PyTorch implementation of GNINA scoring function’s Caffe training script.

gninatorch.training.options(args: List[str] | None = None)[source]

Define options and parse arguments.

Parameters:: args (Optional[List[str]]) – List of command line arguments

gninatorch.training.training(args)[source]

Main function for training GNINA scoring function.

Parameters:: args – Command line arguments

Notes

Training might start off slow because the molgrid.ExampleProvider is caching the structures that are read from .gninatypes files. The training then speeds up considerably.