machine learning - What is "epoch" in keras.models.Model.fit?

Question

What is "epoch" in keras.models.Model.fit? Is it one gradient update? If it is more than one gradient update, then what is defining an epoch?

Suppose I am feeding my own batches to fit. I would regard "epoch" as finishing to process entire training set (is this correct)? Then how to control keras for this way? Can I set batch_size equal to x and y size and epochs to 1?

question from:https://stackoverflow.com/questions/44907377/what-is-epoch-in-keras-models-model-fit

Answer 1

Here is how Keras documentation defines an epoch:

Epoch: an arbitrary cutoff, generally defined as "one pass over the entire dataset", used to separate training into distinct phases, which is useful for logging and periodic evaluation.

So, in other words, a number of epochs means how many times you go through your training set.

The model is updated each time a batch is processed, which means that it can be updated multiple times during one epoch. If batch_size is set equal to the length of x, then the model will be updated once per epoch.