A fully customizable neural network from scratch in numpy.It can be made of any depth and each layer can have any number of neurons. It has adam optimiser with mini batch gradient descent. It also has certain tools for checking performance of neural network like gradient checking etc. L2 regularization is used which is too optional.
kaif3007/Neural_Network_from_scratch
Folders and files
| Name | Name | Last commit date | ||
|---|---|---|---|---|