Matches in SemOpenAlex for { <https://semopenalex.org/work/W4281623710> ?p ?o ?g. }
Showing items 1 to 74 of
74
with 100 items per page.
- W4281623710 abstract "The process of optimizing the latency of DNN operators with ML models and hardware-in-the-loop, called auto-tuning, has established itself as a pervasive method for the deployment of neural networks. From a search space of loop-optimizations, the candidate providing the best performance has to be selected. Performance of individual configurations is evaluated through hardware measurements. The combinatorial explosion of possible configurations, together with the cost of hardware evaluation makes exhaustive explorations of the search space infeasible in practice. Machine Learning methods, like random forests or reinforcement learning are used to aid in the selection of candidates for hardware evaluation. For general purpose hardware like x86 and GPGPU architectures impressive performance gains can be achieved, compared to hand-optimized libraries like cuDNN. The method is also useful in the space of hardware accelerators with less wide-spread adoption, where a high-performance library is not always available. However, hardware accelerators are often less flexible with respect to their programming which leads to operator configurations not executable on the hardware target. This work evaluates how these invalid configurations affect the auto-tuning process and its underlying performance prediction model for the VTA hardware. From these results, a validity-driven initialization method for AutoTVM is developed, only requiring 41.6% of the necessary hardware measurements to find the best solution, while improving search robustness." @default.
- W4281623710 created "2022-06-13" @default.
- W4281623710 creator A5014251134 @default.
- W4281623710 creator A5053327260 @default.
- W4281623710 creator A5074802358 @default.
- W4281623710 creator A5075535959 @default.
- W4281623710 date "2022-05-31" @default.
- W4281623710 modified "2023-09-28" @default.
- W4281623710 title "HW-Aware Initialization of DNN Auto-Tuning to Improve Exploration Time and Robustness" @default.
- W4281623710 doi "https://doi.org/10.48550/arxiv.2205.15568" @default.
- W4281623710 hasPublicationYear "2022" @default.
- W4281623710 type Work @default.
- W4281623710 citedByCount "0" @default.
- W4281623710 crossrefType "posted-content" @default.
- W4281623710 hasAuthorship W4281623710A5014251134 @default.
- W4281623710 hasAuthorship W4281623710A5053327260 @default.
- W4281623710 hasAuthorship W4281623710A5074802358 @default.
- W4281623710 hasAuthorship W4281623710A5075535959 @default.
- W4281623710 hasBestOaLocation W42816237101 @default.
- W4281623710 hasConcept C104317684 @default.
- W4281623710 hasConcept C111919701 @default.
- W4281623710 hasConcept C113775141 @default.
- W4281623710 hasConcept C114466953 @default.
- W4281623710 hasConcept C13164978 @default.
- W4281623710 hasConcept C154945302 @default.
- W4281623710 hasConcept C160145156 @default.
- W4281623710 hasConcept C170723468 @default.
- W4281623710 hasConcept C173608175 @default.
- W4281623710 hasConcept C185592680 @default.
- W4281623710 hasConcept C199360897 @default.
- W4281623710 hasConcept C2777904410 @default.
- W4281623710 hasConcept C41008148 @default.
- W4281623710 hasConcept C42935608 @default.
- W4281623710 hasConcept C50644808 @default.
- W4281623710 hasConcept C55493867 @default.
- W4281623710 hasConcept C63479239 @default.
- W4281623710 hasConcept C9390403 @default.
- W4281623710 hasConcept C98045186 @default.
- W4281623710 hasConceptScore W4281623710C104317684 @default.
- W4281623710 hasConceptScore W4281623710C111919701 @default.
- W4281623710 hasConceptScore W4281623710C113775141 @default.
- W4281623710 hasConceptScore W4281623710C114466953 @default.
- W4281623710 hasConceptScore W4281623710C13164978 @default.
- W4281623710 hasConceptScore W4281623710C154945302 @default.
- W4281623710 hasConceptScore W4281623710C160145156 @default.
- W4281623710 hasConceptScore W4281623710C170723468 @default.
- W4281623710 hasConceptScore W4281623710C173608175 @default.
- W4281623710 hasConceptScore W4281623710C185592680 @default.
- W4281623710 hasConceptScore W4281623710C199360897 @default.
- W4281623710 hasConceptScore W4281623710C2777904410 @default.
- W4281623710 hasConceptScore W4281623710C41008148 @default.
- W4281623710 hasConceptScore W4281623710C42935608 @default.
- W4281623710 hasConceptScore W4281623710C50644808 @default.
- W4281623710 hasConceptScore W4281623710C55493867 @default.
- W4281623710 hasConceptScore W4281623710C63479239 @default.
- W4281623710 hasConceptScore W4281623710C9390403 @default.
- W4281623710 hasConceptScore W4281623710C98045186 @default.
- W4281623710 hasLocation W42816237101 @default.
- W4281623710 hasLocation W42816237102 @default.
- W4281623710 hasOpenAccess W4281623710 @default.
- W4281623710 hasPrimaryLocation W42816237101 @default.
- W4281623710 hasRelatedWork W2008876287 @default.
- W4281623710 hasRelatedWork W2095112382 @default.
- W4281623710 hasRelatedWork W2107563064 @default.
- W4281623710 hasRelatedWork W2110615297 @default.
- W4281623710 hasRelatedWork W2183042904 @default.
- W4281623710 hasRelatedWork W2295680811 @default.
- W4281623710 hasRelatedWork W2351404747 @default.
- W4281623710 hasRelatedWork W3035319544 @default.
- W4281623710 hasRelatedWork W3038662377 @default.
- W4281623710 hasRelatedWork W4310095167 @default.
- W4281623710 isParatext "false" @default.
- W4281623710 isRetracted "false" @default.
- W4281623710 workType "article" @default.