A callback is basically a Keras library function that can interact with our model during the training session to check on its internal state and save relevant training statistics for later scrutiny. While quite a few callback functions exist in keras.callbacks, we will introduce a few that are crucial. For those of you who are more technically oriented, Keras even lets you construct custom callbacks. To use a callback, you simply pass it to the fit parameter using the keyword argument callbacks. Note that the history callback is automatically applied to every Keras model, and so it does not need to be specified as long as you define the fitting process as a variable. This lets you recover the associated history object.
Importantly, if you initiated a training session previously in your Jupyter Notebook, then calling the fit() parameter on the model will continue training...
