Matches in SemOpenAlex for { <https://semopenalex.org/work/W3165698711> ?p ?o ?g. }
Showing items 1 to 96 of
96
with 100 items per page.
- W3165698711 abstract "With the recent trend of on-device deep learning, inference latency has become a crucial metric in running Deep Neural Network (DNN) models on various mobile and edge devices. To this end, latency prediction of DNN model inference is highly desirable for many tasks where measuring the latency on real devices is infeasible or too costly, such as searching for efficient DNN models with latency constraints from a huge model-design space. Yet it is very challenging and existing approaches fail to achieve a high accuracy of prediction, due to the varying model-inference latency caused by the runtime optimizations on diverse edge devices. In this paper, we propose and develop nn-Meter, a novel and efficient system to accurately predict the inference latency of DNN models on diverse edge devices. The key idea of nn-Meter is dividing a whole model inference into kernels, i.e., the execution units on a device, and conducting kernel-level prediction. nn-Meter builds atop two key techniques: (i) kernel detection to automatically detect the execution unit of model inference via a set of well-designed test cases; and (ii) adaptive sampling to efficiently sample the most beneficial configurations from a large space to build accurate kernel-level latency predictors. Implemented on three popular platforms of edge hardware (mobile CPU, mobile GPU, and Intel VPU) and evaluated using a large dataset of 26,000 models, nn-Meter significantly outperforms the prior state-of-the-art." @default.
- W3165698711 created "2021-06-07" @default.
- W3165698711 creator A5012503019 @default.
- W3165698711 creator A5018118687 @default.
- W3165698711 creator A5027782298 @default.
- W3165698711 creator A5047883479 @default.
- W3165698711 creator A5062955154 @default.
- W3165698711 creator A5071351577 @default.
- W3165698711 creator A5079830250 @default.
- W3165698711 date "2021-06-24" @default.
- W3165698711 modified "2023-10-04" @default.
- W3165698711 title "nn-Meter" @default.
- W3165698711 cites W2147657366 @default.
- W3165698711 cites W2903650079 @default.
- W3165698711 cites W2911964244 @default.
- W3165698711 cites W2961619211 @default.
- W3165698711 cites W2963125010 @default.
- W3165698711 cites W2963918968 @default.
- W3165698711 cites W2967733054 @default.
- W3165698711 cites W2980137827 @default.
- W3165698711 cites W3035130950 @default.
- W3165698711 cites W3102476541 @default.
- W3165698711 cites W3102510044 @default.
- W3165698711 doi "https://doi.org/10.1145/3458864.3467882" @default.
- W3165698711 hasPublicationYear "2021" @default.
- W3165698711 type Work @default.
- W3165698711 sameAs 3165698711 @default.
- W3165698711 citedByCount "40" @default.
- W3165698711 countsByYear W31656987112021 @default.
- W3165698711 countsByYear W31656987112022 @default.
- W3165698711 countsByYear W31656987112023 @default.
- W3165698711 crossrefType "proceedings-article" @default.
- W3165698711 hasAuthorship W3165698711A5012503019 @default.
- W3165698711 hasAuthorship W3165698711A5018118687 @default.
- W3165698711 hasAuthorship W3165698711A5027782298 @default.
- W3165698711 hasAuthorship W3165698711A5047883479 @default.
- W3165698711 hasAuthorship W3165698711A5062955154 @default.
- W3165698711 hasAuthorship W3165698711A5071351577 @default.
- W3165698711 hasAuthorship W3165698711A5079830250 @default.
- W3165698711 hasConcept C108583219 @default.
- W3165698711 hasConcept C111919701 @default.
- W3165698711 hasConcept C113775141 @default.
- W3165698711 hasConcept C114614502 @default.
- W3165698711 hasConcept C119857082 @default.
- W3165698711 hasConcept C127413603 @default.
- W3165698711 hasConcept C138236772 @default.
- W3165698711 hasConcept C154945302 @default.
- W3165698711 hasConcept C162307627 @default.
- W3165698711 hasConcept C176217482 @default.
- W3165698711 hasConcept C186967261 @default.
- W3165698711 hasConcept C21547014 @default.
- W3165698711 hasConcept C2776214188 @default.
- W3165698711 hasConcept C33923547 @default.
- W3165698711 hasConcept C41008148 @default.
- W3165698711 hasConcept C50644808 @default.
- W3165698711 hasConcept C74193536 @default.
- W3165698711 hasConcept C76155785 @default.
- W3165698711 hasConcept C79974875 @default.
- W3165698711 hasConcept C82876162 @default.
- W3165698711 hasConceptScore W3165698711C108583219 @default.
- W3165698711 hasConceptScore W3165698711C111919701 @default.
- W3165698711 hasConceptScore W3165698711C113775141 @default.
- W3165698711 hasConceptScore W3165698711C114614502 @default.
- W3165698711 hasConceptScore W3165698711C119857082 @default.
- W3165698711 hasConceptScore W3165698711C127413603 @default.
- W3165698711 hasConceptScore W3165698711C138236772 @default.
- W3165698711 hasConceptScore W3165698711C154945302 @default.
- W3165698711 hasConceptScore W3165698711C162307627 @default.
- W3165698711 hasConceptScore W3165698711C176217482 @default.
- W3165698711 hasConceptScore W3165698711C186967261 @default.
- W3165698711 hasConceptScore W3165698711C21547014 @default.
- W3165698711 hasConceptScore W3165698711C2776214188 @default.
- W3165698711 hasConceptScore W3165698711C33923547 @default.
- W3165698711 hasConceptScore W3165698711C41008148 @default.
- W3165698711 hasConceptScore W3165698711C50644808 @default.
- W3165698711 hasConceptScore W3165698711C74193536 @default.
- W3165698711 hasConceptScore W3165698711C76155785 @default.
- W3165698711 hasConceptScore W3165698711C79974875 @default.
- W3165698711 hasConceptScore W3165698711C82876162 @default.
- W3165698711 hasLocation W31656987111 @default.
- W3165698711 hasOpenAccess W3165698711 @default.
- W3165698711 hasPrimaryLocation W31656987111 @default.
- W3165698711 hasRelatedWork W10172392 @default.
- W3165698711 hasRelatedWork W12275997 @default.
- W3165698711 hasRelatedWork W13854172 @default.
- W3165698711 hasRelatedWork W2288985 @default.
- W3165698711 hasRelatedWork W2724962 @default.
- W3165698711 hasRelatedWork W3313735 @default.
- W3165698711 hasRelatedWork W5469727 @default.
- W3165698711 hasRelatedWork W6520677 @default.
- W3165698711 hasRelatedWork W8834286 @default.
- W3165698711 hasRelatedWork W9654553 @default.
- W3165698711 isParatext "false" @default.
- W3165698711 isRetracted "false" @default.
- W3165698711 magId "3165698711" @default.
- W3165698711 workType "article" @default.