molprop.training.train_model

molprop.training.train_model(args, model, datamodule)

Performs the training process, i.e. initializing optimizer and early stopping, for example, looping over epochs and getting the evaluation metrics.

Parameters:

• args – Parameters from defaultconfig.yaml.

• model – The model used for the training process.

• datamodule – Splits the dataset into three sets: training, validation, test.

Returns:

Best model of the training process, if the early stopping or max. number of epochs is reached. NoneType: Otherwise.

Return type:

torch.nn.Module