A neural netwrok with a single hidden layer coded from scratch (using only Numpy and xlogy from scipy). With this exercise, it's easy to see how fully-connected neural networks with several layers are built. This neural network implements mini-batch stochastic gradient descent. It is applied to the Fashion MNIST dataset.
mghrear/Numpy-NN
Folders and files
| Name | Name | Last commit date | ||
|---|---|---|---|---|