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

Traditional deep learning compilers rely on heuristics for subgraphgeneration, which impose extra constraints on graph optimization, e.g., eachsubgraph can only contain at most one complex operator. In this paper, wepropose AGO, a framework for graph optimization with arbitrary structures toboost the inference performance of deep models by removing such constraints. Tocreate new optimization opportunities for complicated subgraphs, we proposeintensive operator fusion, which can effectively stitch multiple complexoperators together for better performance. Further, we design a graphpartitioning scheme that allows an arbitrary structure for each subgraph whileguaranteeing the acyclic property among all generated subgraphs. Additionally,to enable efficient performance tuning on complicated subgraphs, we devise anovel divide-and-conquer tuning mechanism to orchestrate different systemcomponents. Through extensive experiments on various neural networks and mobiledevices, we show that our system can improve the inference performance by up to3.3x when compared with state-of-the-art deep compilers.

AGO: Boosting Mobile AI Inference Performance by Removing Constraints on Graph Optimization