callbacks

Description

Callback functionality and typical callbacks for training. A callback simply is a mapping from a CallbackEvents instance to a callable that accepts the callback context. The run_callbacks() method can be used to run those callables associated with a given event from a list of callbacks. See Callback for details.

Classes

Callback

A callback base class that eases implementing a custom callback handle.

CallbackEvents

Supported callback events for training and evaluation.

CsvLoggingCallback

Extract the values stored in matplotlib figures and store them as CSV.

LoggingCallback

Log batch and epoch KPI results.

ProgressBarUpdater

Update the progress bar postfix after each batch and epoch.

TensorboardLogger

Write batch and epoch KPI results to a tensorboard log directory.

Functions

run_callbacks(callbacks, event[, ...])

Run the given callback collection in order.