A neural network implementation written from scratch (i.e. no ML library is used) that graphically displays the network structure, weights, and biases while it is training.
You can configure any network topology, but convolutional networks are not displayed well graphically.
Run 'TestNeuralNet' for a demo. See MainWindow.xaml.cs in the TestNeuralNet project for an example of how to train and use it. There are several different data sets coded in that file that you can train for--just uncomment the training set you want to experiment with.
NeuralNet training is multithreaded in CPU (not GPU).