1 paper · 2 filters
Santiago Miret, Vui Seng Chua, Mattias Marder +3
Mixed-precision quantization is a powerful tool to enable memory and compute savings of neural network workloads by deploying different sets of bit-width precisions on separate com…