A Fixed-Point Quantization Technique for Convolutional Neural Networks Based on Weight Scaling.

Norbert Mitschke, Michael Heizmann, Klaus-Henning Noffz, Ralf Wittmann

Browse the full ICIP paper archive.