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 of deep neural networks (DNN) has become a key element in theefforts of embedding such networks on end-user devices. However, currentquantization methods usually suffer from costly accuracy degradation. In thispaper, we propose a new method for Enhanced Post Training Quantization namedEPTQ. The method is based on knowledge distillation with an adaptive weightingof layers. In addition, we introduce a new label-free technique forapproximating the Hessian trace of the task loss, named Label-Free Hessian.This technique removes the requirement of a labeled dataset for computing theHessian. The adaptive knowledge distillation uses the Label-Free Hessiantechnique to give greater attention to the sensitive parts of the model whileperforming the optimization. Empirically, by employing EPTQ we achievestate-of-the-art results on a wide variety of models, tasks, and datasets,including ImageNet classification, COCO object detection, and Pascal-VOC forsemantic segmentation. We demonstrate the performance and compatibility of EPTQon an extended set of architectures, including CNNs, Transformers, hybrid, andMLP-only models.

EPTQ: Enhanced Post-Training Quantization via Label-Free Hessian