Trainers

Main trainers

Trainer(model, criterion, data_loader[, ...])	Trainer base class.
TrainerGrad(model, criterion, data_loader, ...)	The default gradient descent trainer.
TrainerEmbedding(model, criterion, ...[, ...])	An (unsupervised) trainer that transforms input data into linearly-separable embedding vectors that form clusters.
TrainerAutoencoder(model, criterion, ...[, ...])	An unsupervised AutoEncoder trainer that not only transforms inputs to meaningful embeddings but also aims to restore the input signal from it.

Idle test “trainer”

Test(model, criterion, data_loader[, ...])

A Test trainer that only evaluates a model.

Interpretable mask trainers

MaskTrainer(image_shape[, accuracy_measure, ...])

Interpretable Explanations of Black Boxes by Meaningful Perturbation [R24f8eaaaa2d9-1].