It's a machine learning example made with a Deep Feedfoward Neural Network using MNIST Dataset
Train New Model: Train a model from zero and watch it learn with the images. You can also predict while it's learning.
Load Trained Model: Load a model already trained and have fun with the predictions.
Predict Number or Spacebar: Predict number with the test data and see the prediction in the top of the screen.
Stop Training: Stop the training of your model.
Continue Training: Continue the training of your model.
The Neural Network it's fed with the images of MNIST and try to guess which number is it.
For every error that it does, it tries to correct itself to come the closest as possible of the real output, in our case, the digit.
After some training, the Neural Network learns how identify the human digits by itself, and predict the result with high accuracy. The results could be seen in the right-side of the screen, and you can play around with it =D
Pull requests are welcome. For major changes, please open an issue first to discuss what you would like to change.