For now, this is a simple demo where a Flask server exposes an API to publish and consume events in the form of JSON objects. The Keras callback RemoteMonitor
can publish events to the server, and the Hualos landing page listens to the server and displays incoming data on a c3.js graph.
Example:
python api.py
http://localhost:9000/
RemoteMonitor
callback:from keras import callbacks
remote = callbacks.RemoteMonitor(root='http://localhost:9000')
model.fit(X_train, Y_train, batch_size=batch_size, nb_epoch=nb_epoch, validation_data=(X_test, Y_test), callbacks=[remote])
Python:
JS (included in the repo):