A toolkit to optimize ML models for deployment for Keras and TensorFlow, including quantization and pruning.