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

We introduce a data-free quantization method for deep neural networks thatdoes not require fine-tuning or hyperparameter selection. It achievesnear-original model performance on common computer vision architectures andtasks. 8-bit fixed-point quantization is essential for efficient inference onmodern deep learning hardware. However, quantizing models to run in 8-bit is anon-trivial task, frequently leading to either significant performancereduction or engineering time spent on training a network to be amenable toquantization. Our approach relies on equalizing the weight ranges in thenetwork by making use of a scale-equivariance property of activation functions.In addition the method corrects biases in the error that are introduced duringquantization. This improves quantization accuracy performance, and can beapplied to many common computer vision architectures with a straight forwardAPI call. For common architectures, such as the MobileNet family, we achievestate-of-the-art quantized model performance. We further show that the methodalso extends to other computer vision architectures and tasks such as semanticsegmentation and object detection.

Data-Free Quantization Through Weight Equalization and Bias Correction