In this episode I speak about neural network quantization, a technique that makes networks feasible for embedded systems and small devices.