English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Quantization is an effective method for reducing memory footprint andinference time of Neural Networks, e.g., for efficient inference in the cloud,especially at the edge. However, ultra low precision quantization could lead tosignificant degradation in model generalization. A promising method to addressthis is to perform mixed-precision quantization, where more sensitive layersare kept at higher precision. However, the search space for a mixed-precisionquantization is exponential in the number of layers. Recent work has proposedHAWQ, a novel Hessian based framework, with the aim of reducing thisexponential search space by using second-order information. While promising,this prior work has three major limitations: (i) HAWQV1 only uses the topHessian eigenvalue as a measure of sensitivity and do not consider the rest ofthe Hessian spectrum; (ii) HAWQV1 approach only provides relative sensitivityof different layers and therefore requires a manual selection of themixed-precision setting; and (iii) HAWQV1 does not consider mixed-precisionactivation quantization. Here, we present HAWQV2 which addresses theseshortcomings. For (i), we perform a theoretical analysis showing that a bettersensitivity metric is to compute the average of all of the Hessian eigenvalues.For (ii), we develop a Pareto frontier based method for selecting the exact bitprecision of different layers without any manual selection. For (iii), weextend the Hessian analysis to mixed-precision activation quantization. We havefound this to be very beneficial for object detection. We show that HAWQV2achieves new state-of-the-art results for a wide range of tasks.

HAWQ-V2: Hessian Aware trace-Weighted Quantization of Neural Networks