PRESS 1 AND 2 TO SEE NETWORK INFO ------------------------------------------------------------------------- WHAT IT CAN DO ------------------------------------------------------------------------- Supports Any Size Network: - Any number of layers - Any number of neurones in each layer Visually Represents Network - Links correspond to strength of weights - Neurons light up according to output Calculates the Forward Pass Through the Network - Has a ReLU Activation Function for hidden layers - Uses SoftMax for confidence values on output THIS PROJECT IS INCOMPLETE I will try to turn it into a digit recogniser Currently the network is processing 10,000 images from the MNIST data set which I imported It is simply calculating an output however and there is no learning and backpropagation happening yet https://turbowarp.org/1296785493
