Deloitte interview question

What are the differences between stochastic gradient descent and mini-batch gradient descent?