Given a trained network stored in JSON-format, this feature can be used to monitor in a visually appealing way the probability distribution of the features being modelled. It can be used for debugging purposes but also just to play around with an existing model.

Here is an example of how it is used. In this example I have hit the "Sample next" button to generate each character.

Steps needed to be taken:

1. Train a model
2. Stop the training process when the model is sufficiently trained. This can be done either by hitting CTRL-C. When the program is stopped the network weights are automatically stored. You can also store a network during training, periodically, by starting the program with the '-st' argument.
3. Load the JSON-file into the GUI
4. Either sample a character one-by-one using Sample next
5. .. or: Give it a sequence of characters to start it off

Include the .json file in the "models" folder to analyze the behaviour of an already trained model, having read the first Harry Potter book many times.