Faster Ternary and Binary Neural Network Inference on CPU by Reducing Popcount Overhead.

Olivier Fischer, Shien Zhu, Gustavo Alonso

Browse the full ISLPED paper archive.