tfm.core.train_utils.create_trainer

Create trainer.

tfm.core.train_utils.create_trainer(
    params: tfm.core.base_trainer.ExperimentConfig,
    task: tfm.core.base_task.Task,
    train: bool,
    evaluate: bool,
    checkpoint_exporter: Optional[tfm.core.train_utils.BestCheckpointExporter] = None,
    trainer_cls=tfm.core.base_trainer.Trainer
) -> tfm.core.base_trainer.Trainer