I am new to Keras and am trying to customize my training step in Keras.
More specifically, in the stochastic gradient descent algorithm, after feedforwarding a mini-b
