does output from keras fit show average of loss?

Question

I am using binary cross entropy, and I have 2 epochs:

batch_size = 32
epochs = 2
History = model.fit(padded_train, y_train, batch_size = batch_size, epochs = epochs, validation_split = 0.1)

Now i get the following output

loss: 0.0771 - accuracy: 0.9763 - val_loss: 0.0575 - val_accuracy: 0.9806

Now what i am wondering is, whether the loss i am seeing here signifies an average of the loss for all observations, or maybe just the last one?

Answer 1

This is only for the 2nd epoch you should get an output below saying: Accuracy x%. x is representing the accuracy of the whole 2 epochs