espnet2.train.uasr_trainer.UASRTrainer

source

class espnet2.train.uasr_trainer.UASRTrainer

Bases: Trainer

Trainer for GAN-based UASR training.

If you’d like to use this trainer, the model must inherit espnet.train.abs_gan_espnet_model.AbsGANESPnetModel.

classmethod add_arguments(parser: ArgumentParser)

Add additional arguments for GAN-trainer.

classmethod build_options(args: Namespace) → TrainerOptions

Build options consumed by train(), eval(), and plot_attention().

classmethod train_one_epoch(model: Module, iterator: Iterable[Tuple[List[str], Dict[str, Tensor]]], optimizers: Sequence[Optimizer], schedulers: Sequence[AbsScheduler | None], scaler: GradScaler | None, reporter: SubReporter, summary_writer, options: UASRTrainerOptions, distributed_option: DistributedOption) → bool

Train one epoch for UASR.

classmethod validate_one_epoch(model: Module, iterator: Iterable[Dict[str, Tensor]], reporter: SubReporter, options: UASRTrainerOptions, distributed_option: DistributedOption) → None

Validate one epoch.