Usual gradient descent will get trapped at an area minimum rather then a global minimal, resulting in a subpar community. In ordinary gradient descent, we get all our rows and plug them into your exact same neural network, Consider the weights, and after that change them. In wide conditions,