Stop the training and execute the next lines of code

[Rhuax]

Wouldn't be good to decide when to stop the execution of the fit() method during a training and execute the next lines? Or it would be bad in terms of performance since it needs to check every time if a specific key is pressed?
I think it would be good to stop training when a reasonable good loss value is seen.

[NasenSpray]

You can use callbacks to stop training at any point, e.g. see EarlyStopping