sigmoid, tanh and relu should be moved to neural_network/activation_functions from maths/

[AdarshAcharya5]

Feature description

sigmoid.py, tanh.py, sigmoid_linear_unit.py and relu.py should be moved to neural_network/activation_functions.

[rohan472000]

@AdarshAcharya5 , why you want to move those .py files?

[anchor54]

is this still a requirement and open for contribution?

[tianyizheng02]

@rohan472000 Functions such as sigmoid, tanh, and ReLU are nowadays commonly used as activation functions for neural networks. Sigmoid (or more specifically the logistic function) and tanh are still well-known mathematical functions outside of ML, so I think those should stay in the maths/ directory. However, ReLU is pretty much only known as an activation function as far as I know, so I think it's appropriate to move that algorithm to neural_network/activation_functions.

[piyush-poddar]

Hi I have raised a PR, please review it.