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

Conventional model quantization methods use a fixed quantization scheme todifferent data samples, which ignores the inherent "recognition difficulty"differences between various samples. We propose to feed different data sampleswith varying quantization schemes to achieve a data-dependent dynamicinference, at a fine-grained layer level. However, enabling this adaptiveinference with changeable layer-wise quantization schemes is challengingbecause the combination of bit-widths and layers is growing exponentially,making it extremely difficult to train a single model in such a vast searchingspace and use it in practice. To solve this problem, we present the ArbitraryBit-width Network (ABN), where the bit-widths of a single deep network canchange at runtime for different data samples, with a layer-wise granularity.Specifically, first we build a weight-shared layer-wise quantizable"super-network" in which each layer can be allocated with multiple bit-widthsand thus quantized differently on demand. The super-network provides aconsiderably large number of combinations of bit-widths and layers, each ofwhich can be used during inference without retraining or storing myriad models.Second, based on the well-trained super-network, each layer's runtime bit-widthselection decision is modeled as a Markov Decision Process (MDP) and solved byan adaptive inference strategy accordingly. Experiments show that thesuper-network can be built without accuracy degradation, and the bit-widthsallocation of each layer can be adjusted to deal with various inputs on thefly. On ImageNet classification, we achieve 1.1% top1 accuracy improvementwhile saving 36.2% BitOps.

Arbitrary Bit-width Network: A Joint Layer-Wise Quantization and Adaptive Inference Approach