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

Achieving efficient execution of machine learning models has attractedsignificant attention recently. To generate tensor programs efficiently, a keycomponent of DNN compilers is the cost model that can predict the performanceof each configuration on specific devices. However, due to the rapid emergenceof hardware platforms, it is increasingly labor-intensive to traindomain-specific predictors for every new platform. Besides, current design ofcost models cannot provide transferable features between different hardwareaccelerators efficiently and effectively. In this paper, we propose Moses, asimple and efficient design based on the lottery ticket hypothesis, which fullytakes advantage of the features transferable to the target device via domainadaptation. Compared with state-of-the-art approaches, Moses achieves up to1.53X efficiency gain in the search stage and 1.41X inference speedup onchallenging DNN benchmarks.

Moses: Efficient Exploitation of Cross-device Transferable Features for Tensor Program Optimization