TrainProgressMonitor

class torchtnt.framework.callbacks.TrainProgressMonitor(loggers: Union[MetricLogger, List[MetricLogger]])

A callback which logs training progress in terms of steps vs epochs. This is helpful to visualize when the end of data occurs across epochs, especially for iterable datasets. This callback writes to the logger at the beginning of training, and at the end of every epoch.

Parameters:loggers – Either a torchtnt.loggers.logger.MetricLogger or list of torchtnt.loggers.logger.MetricLogger

on_train_epoch_end(state: State, unit: TrainUnit[TTrainData]) → None

Hook called after a train epoch ends.

on_train_start(state: State, unit: TrainUnit[TTrainData]) → None

Hook called before training starts.