35. Batch vs Stochastic Gradient Descent

Stochastic Gradient Descent in Keras

Stochastic gradient descent is very easy to implement in Keras. All we need to do is specify the size of the batches in the training process, as follows:

model.fit(X_train, y_train, epochs=1000, batch_size=100, verbose=0)

In here, we are splitting our data in batches of 100.