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

With the growing demand for vision applications and deployment across edgedevices, the development of hardware-friendly architectures that maintainperformance during device deployment becomes crucial. Neural architecturesearch (NAS) techniques explore various approaches to discover efficientarchitectures for diverse learning tasks in a computationally efficient manner.In this paper, we present the next-generation neural architecture design forcomputationally efficient neural architecture distillation - DONNAv2 .Conventional NAS algorithms rely on a computationally extensive stage where anaccuracy predictor is learned to estimate model performance within searchspace. This building of accuracy predictors helps them predict the performanceof models that are not being finetuned. Here, we have developed an elegantapproach to eliminate building the accuracy predictor and extend DONNA to acomputationally efficient setting. The loss metric of individual blocks formingthe network serves as the surrogate performance measure for the sampled modelsin the NAS search stage. To validate the performance of DONNAv2 we haveperformed extensive experiments involving a range of diverse vision tasksincluding classification, object detection, image denoising, super-resolution,and panoptic perception network (YOLOP). The hardware-in-the-loop experimentswere carried out using the Samsung Galaxy S10 mobile platform. Notably, DONNAv2reduces the computational cost of DONNA by 10x for the larger datasets.Furthermore, to improve the quality of NAS search space, DONNAv2 leverages ablock knowledge distillation filter to remove blocks with high inference costs.

DONNAv2 -- Lightweight Neural Architecture Search for Vision tasks