* take_picture() and display_picture() now take a single input pattern, or list of patterns, 1 per input bank * wrapped Model.compile() with Network.compile() to allow metrics=["tolerance_accuracy"]
0.0.7
Major changes:
* colormap is now (colormap_name, min, max) * added get_tolerance(), set_tolerance() for "tolerance_accuracy" metric
0.0.6
Major changes:
* add accuracy metric to track in SimpleNetwork
0.0.5
Major changes:
* allow plotting to work with any accuracy metric * plot a final plot at end * display an ending message showing how many epochs, final metrics
* Stop training when: * monitored metric stalls * loss or val_loss falls below a value * acc or accuracy is above a value * SimpleNetwork * display_colormap * predict * set_config_layers * rotate and scale not sticky