nntoolbox.callbacks.checkpoint module

class nntoolbox.callbacks.checkpoint.ModelCheckpoint(learner, filepath: str, monitor: str = 'loss', save_best_only: bool = True, mode: str = 'min', period: int = 1)

Bases: nntoolbox.callbacks.callbacks.Callback

on_epoch_end(logs: Dict[str, Any]) → bool

class nntoolbox.callbacks.checkpoint.OptimizerCheckPoint(filepath: str, monitor: str = 'loss', save_best_only: bool = True, mode: str = 'min', period: int = 1)

Bases: nntoolbox.callbacks.callbacks.Callback

on_epoch_end(logs: Dict[str, Any]) → bool

class nntoolbox.callbacks.checkpoint.ResumeFromCheckpoint(model_path: Optional[str] = None, optimizer_path: Optional[str] = None)

Bases: nntoolbox.callbacks.callbacks.Callback

Resume from previous checkpoint

on_train_begin()