Visualizer requires conditions in model

[chadcwilliams]

When loading a model using the visualizer, it relies on the conditions parameter provided to the model during training. This does not work with the autoencoder, which does not take a conditions parameter.