nlgm.train Module

nlgm.train.train_and_evaluate(model, train_loader, test_loader, epochs=10, device=device(type='cpu'), progress_bar=True)

Train a model and evaluate on a test loader each epoch.

Parameters:

modelnn.Module

Model to train and evaluate.

train_loadertorch.utils.data.DataLoader

Training data loader.

test_loadertorch.utils.data.DataLoader

Test data loader.

epochsint, default=10

Number of training epochs.

devicetorch.device, default=torch.device(“cpu”)

Device used for training and evaluation.

progress_barbool, default=True

Whether to show epoch-level progress.

Returns:

list[float]

Per-epoch training losses.

float

Final epoch test loss.