Technical Report: NEMO DNN Quantization for Deployment Model

Abstract

This technical report aims at defining a formal framework for Deep Neural Network (DNN) layer-wise quantization, focusing in particular on the problems related to the final deployment. It also acts as a documentation for the NEMO (NEural Minimization for pytOrch) framework. It describes the four DNN representations used in NEMO (FullPrecision, FakeQuantized, QuantizedDeployable and IntegerDeployable), focusing in particular on a formal definition of the latter two. An important feature of this model, and in particular the IntegerDeployable representation, is that it enables DNN inference using purely integers - without resorting to real-valued numbers in any part of the computation and without relying on an explicit fixed-point numerical representation.

12 pages, technical report

References in corpus (3)

• PACT: Parameterized Clipping Activation for Quantized Neural Networks
• Memory-Driven Mixed Low Precision Quantization For Enabling Deep Network Inference On Microcontrollers
• Additive Noise Annealing and Approximation Properties of Quantized Neural Networks

Cited by in corpus (4)

• A TinyML Platform for On-Device Continual Learning with Quantized Latent Replays
• Ultra-low Power Deep Learning-based Monocular Relative Localization Onboard Nano-quadrotors
• Deep Neural Network Architecture Search for Accurate Visual Pose Estimation aboard Nano-UAVs
• Utilizing Explainable AI for Quantization and Pruning of Deep Neural Networks