Save the model at the beginning/end of training

custom_modelcheckpoint.py

class CustomModelCheckpoint(ModelCheckpoint):
    """Custom ModelCheckpoint.
    
    Save the model at the beginning (a random init model) and at the
    end of the training.
    """
    def __init__(self, *args, **kwargs):
        super(CustomModelCheckpoint, self).__init__(*args, **kwargs)
        self.current_epoch = 0
    
    def on_epoch_begin(self, epoch, logs=None):
        self.current_epoch = epoch
        
    def on_train_begin(self, logs=None):
        """Save the initialized model."""
        logs = logs or {}
        epoch = self.current_epoch
        filepath = self.filepath.format(epoch=epoch, **logs)

        if self.verbose > 0:
            print('\nEpoch %05d: saving model to %s' % (epoch + 1, filepath))
        if self.save_weights_only:
            self.model.save_weights(filepath, overwrite=True)
        else:
            self.model.save(filepath, overwrite=True)
            
    def on_train_end(self, logs=None):
        """Save the last model."""
        logs = logs or {}
        epoch = self.current_epoch    # maximum epochs
        filepath = self.filepath.format(epoch=epoch+1, **logs)

        if self.verbose > 0:
            print('\nEpoch %05d: saving model to %s' % (epoch + 1, filepath))
        if self.save_weights_only:
            self.model.save_weights(filepath, overwrite=True)
        else:
            self.model.save(filepath, overwrite=True)

Save the model at the beginning and the END of training.
No need to save the model manually after the training :-)