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

Neural networks have shown great performance in cognitive tasks. Whendeploying network models on mobile devices with limited resources, weightquantization has been widely adopted. Binary quantization obtains the highestcompression but usually results in big accuracy drop. In practice, 8-bit or16-bit quantization is often used aiming at maintaining the same accuracy asthe original 32-bit precision. We observe different layers have differentaccuracy sensitivity of quantization. Thus judiciously selecting differentprecision for different layers/structures can potentially produce moreefficient models compared to traditional quantization methods by striking abetter balance between accuracy and compression rate. In this work, we proposea fine-grained quantization approach for deep neural network compression byrelaxing the search space of quantization bitwidth from discrete to acontinuous domain. The proposed approach applies gradient descend basedoptimization to generate a mixed-precision quantization scheme that outperformsthe accuracy of traditional quantization methods under the same compressionrate.

Differentiable Fine-grained Quantization for Deep Neural Network Compression