Pre-Quantized Deep Learning Models Codified in ONNX to Enable Hardware/Software Co-Design

no code implementations • 4 Oct 2021 • Ulf Hanebutte, Andrew Baldwin, Senad Durakovic, Igor Filipovich, Chien-Chun, Chou, Damian Adamowicz, Derek Chickles, David Hawkes

This paper presents a methodology to separate the quantization process from the hardware-specific model compilation stage via a pre-quantized deep learning model description in standard ONNX format.

Quantization