By converting high-precision data into lower-precision representations quantization makes models more efficient for deployment on AML - View it on GitHub