A fully connected neural network created using numpy only to help intuition and understanding, and tested on MNIST dataset. Can reach ~95% accuracy after relatively few epochs with 784 neurons in first layer, 10 in last and variable amounts in between.
-
Notifications
You must be signed in to change notification settings - Fork 0
IsaacSST/NumPy_Neural_Net
About
A simple feedforward Neural network implemented in numpy and trained on the MNIST dataset
Resources
Stars
Watchers
Forks
Releases
No releases published
Packages 0
No packages published