reducible

An experimental method that optimizes Machine Learning model size without compromising accuracy.