ashpy.callbacks

Callbacks in order to gain control over the training loop.

A callback is a set of functions to be called at given stages of the training procedure. You can use callbacks to implement logging, measure custom metrics or get insight about the training procedure. You can pass a list of callbacks (derived from ashpy.callbacks.callback.Callback) (as the keyword argument callbacks) to the .call() method of the Trainer. The relevant methods of the callbacks will then be called at each stage of the training.

Order:
--on_train_start

----on_epoch_start

------on_batch_start

------on_batch_end

----on_epoch_end

--on_train_end

on_exception – if an Exception was raised

on_event - Called when an event is triggered

The basic class is ashpy.callbacks.callback.Callback . All possible events as listed as Enum inside ashpy.callbacks.events.Event .

Classes

callback.Callback

Callback definition.

counter_callback.CounterCallback

Count events of a specific type.

classifier.LogClassifierCallback

Callback used for logging Classifier images to Tensorboard.

events.Event

Define all possible events.

gan.LogImageGANCallback

Callback used for logging GANs images to Tensorboard.

gan.LogImageGANEncoderCallback

Callback used for logging GANs images to Tensorboard.

save_callback.SaveCallback

Save Callback implementation.

save_callback.SaveFormat

Save Format enum.

save_callback.SaveSubFormat

Save Sub-Format enum.


Modules

callback

Callback definition.

classifier

Classifier callbacks.

counter_callback

Counter Callback implementation.

events

Event definition as Enum.

gan

GAN callbacks.

save_callback

Save weights callback.