Saved in:
Bibliographic Details
Main Authors: Bablani, Deepika, Mckinstry, Jeffrey L., Esser, Steven K., Appuswamy, Rathinakumar, Modha, Dharmendra S.
Format: Preprint
Published: 2023
Subjects:
Online Access:https://arxiv.org/abs/2301.13330
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866929206232875008
author Bablani, Deepika
Mckinstry, Jeffrey L.
Esser, Steven K.
Appuswamy, Rathinakumar
Modha, Dharmendra S.
author_facet Bablani, Deepika
Mckinstry, Jeffrey L.
Esser, Steven K.
Appuswamy, Rathinakumar
Modha, Dharmendra S.
contents For efficient neural network inference, it is desirable to achieve state-of-the-art accuracy with the simplest networks requiring the least computation, memory, and power. Quantizing networks to lower precision is a powerful technique for simplifying networks. As each layer of a network may have different sensitivity to quantization, mixed precision quantization methods selectively tune the precision of individual layers to achieve a minimum drop in task performance (e.g., accuracy). To estimate the impact of layer precision choice on task performance, two methods are introduced: i) Entropy Approximation Guided Layer selection (EAGL) is fast and uses the entropy of the weight distribution, and ii) Accuracy-aware Layer Precision Selection (ALPS) is straightforward and relies on single epoch fine-tuning after layer precision reduction. Using EAGL and ALPS for layer precision selection, full-precision accuracy is recovered with a mix of 4-bit and 2-bit layers for ResNet-50, ResNet-101 and BERT-base transformer networks, demonstrating enhanced performance across the entire accuracy-throughput frontier. The techniques demonstrate better performance than existing techniques in several commensurate comparisons. Notably, this is accomplished with significantly lesser computational time required to reach a solution.
format Preprint
id arxiv_https___arxiv_org_abs_2301_13330
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Efficient and Effective Methods for Mixed Precision Neural Network Quantization for Faster, Energy-efficient Inference
Bablani, Deepika
Mckinstry, Jeffrey L.
Esser, Steven K.
Appuswamy, Rathinakumar
Modha, Dharmendra S.
Machine Learning
Computer Vision and Pattern Recognition
For efficient neural network inference, it is desirable to achieve state-of-the-art accuracy with the simplest networks requiring the least computation, memory, and power. Quantizing networks to lower precision is a powerful technique for simplifying networks. As each layer of a network may have different sensitivity to quantization, mixed precision quantization methods selectively tune the precision of individual layers to achieve a minimum drop in task performance (e.g., accuracy). To estimate the impact of layer precision choice on task performance, two methods are introduced: i) Entropy Approximation Guided Layer selection (EAGL) is fast and uses the entropy of the weight distribution, and ii) Accuracy-aware Layer Precision Selection (ALPS) is straightforward and relies on single epoch fine-tuning after layer precision reduction. Using EAGL and ALPS for layer precision selection, full-precision accuracy is recovered with a mix of 4-bit and 2-bit layers for ResNet-50, ResNet-101 and BERT-base transformer networks, demonstrating enhanced performance across the entire accuracy-throughput frontier. The techniques demonstrate better performance than existing techniques in several commensurate comparisons. Notably, this is accomplished with significantly lesser computational time required to reach a solution.
title Efficient and Effective Methods for Mixed Precision Neural Network Quantization for Faster, Energy-efficient Inference
topic Machine Learning
Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2301.13330