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

Hardware-friendly network quantization (e.g., binary/uniform quantization)can efficiently accelerate the inference and meanwhile reduce memoryconsumption of the deep neural networks, which is crucial for model deploymenton resource-limited devices like mobile phones. However, due to thediscreteness of low-bit quantization, existing quantization methods often facethe unstable training process and severe performance degradation. To addressthis problem, in this paper we propose Differentiable Soft Quantization (DSQ)to bridge the gap between the full-precision and low-bit networks. DSQ canautomatically evolve during training to gradually approximate the standardquantization. Owing to its differentiable property, DSQ can help pursue theaccurate gradients in backward propagation, and reduce the quantization loss inforward process with an appropriate clipping range. Extensive experiments overseveral popular network structures show that training low-bit neural networkswith DSQ can consistently outperform state-of-the-art quantization methods.Besides, our first efficient implementation for deploying 2 to 4-bit DSQ ondevices with ARM architecture achieves up to 1.7$\times$ speed up, comparedwith the open-source 8-bit high-performance inference framework NCNN. [31]

Differentiable Soft Quantization: Bridging Full-Precision and Low-Bit Neural Networks