Matches in SemOpenAlex for { <https://semopenalex.org/work/W3109309915> ?p ?o ?g. }
Showing items 1 to 81 of
81
with 100 items per page.
- W3109309915 endingPage "25" @default.
- W3109309915 startingPage "22" @default.
- W3109309915 abstract "Recently, deep-neural-network (DNN) learning processors for edge devices have been proposed, but they cannot reduce the complexity of over-parameterized network during training. Also, they cannot support energy-efficient zero-skipping because previous methods cannot be performed perfectly in backpropagation and weight gradient update. In this letter, energy-efficient DNN learning processor PNPU is proposed with three key features: 1) stochastic coarse-fine level pruning; 2) adaptive input, output, weight zero skipping; and 3) weight pruning unit with weight sparsity balancer. As a result, PNPU shows 3.14-278.39 TFLOPS/W energy efficiency, at 0.78 V and 50 MHz with FP8 and 0%-90% sparsity condition." @default.
- W3109309915 created "2020-12-07" @default.
- W3109309915 creator A5010104451 @default.
- W3109309915 creator A5015963917 @default.
- W3109309915 creator A5051999681 @default.
- W3109309915 creator A5073016458 @default.
- W3109309915 creator A5077896259 @default.
- W3109309915 creator A5085959296 @default.
- W3109309915 date "2021-01-01" @default.
- W3109309915 modified "2023-09-24" @default.
- W3109309915 title "PNPU: An Energy-Efficient Deep-Neural-Network Learning Processor With Stochastic Coarse–Fine Level Weight Pruning and Adaptive Input/Output/Weight Zero Skipping" @default.
- W3109309915 cites W1980038761 @default.
- W3109309915 cites W2108598243 @default.
- W3109309915 cites W2920954974 @default.
- W3109309915 cites W2921918777 @default.
- W3109309915 cites W2922045004 @default.
- W3109309915 cites W2964282938 @default.
- W3109309915 cites W2965347706 @default.
- W3109309915 cites W3015729306 @default.
- W3109309915 doi "https://doi.org/10.1109/lssc.2020.3041497" @default.
- W3109309915 hasPublicationYear "2021" @default.
- W3109309915 type Work @default.
- W3109309915 sameAs 3109309915 @default.
- W3109309915 citedByCount "7" @default.
- W3109309915 countsByYear W31093099152022 @default.
- W3109309915 countsByYear W31093099152023 @default.
- W3109309915 crossrefType "journal-article" @default.
- W3109309915 hasAuthorship W3109309915A5010104451 @default.
- W3109309915 hasAuthorship W3109309915A5015963917 @default.
- W3109309915 hasAuthorship W3109309915A5051999681 @default.
- W3109309915 hasAuthorship W3109309915A5073016458 @default.
- W3109309915 hasAuthorship W3109309915A5077896259 @default.
- W3109309915 hasAuthorship W3109309915A5085959296 @default.
- W3109309915 hasConcept C105795698 @default.
- W3109309915 hasConcept C108010975 @default.
- W3109309915 hasConcept C108583219 @default.
- W3109309915 hasConcept C11413529 @default.
- W3109309915 hasConcept C154945302 @default.
- W3109309915 hasConcept C155032097 @default.
- W3109309915 hasConcept C165464430 @default.
- W3109309915 hasConcept C186370098 @default.
- W3109309915 hasConcept C33923547 @default.
- W3109309915 hasConcept C41008148 @default.
- W3109309915 hasConcept C50644808 @default.
- W3109309915 hasConcept C6557445 @default.
- W3109309915 hasConcept C86803240 @default.
- W3109309915 hasConceptScore W3109309915C105795698 @default.
- W3109309915 hasConceptScore W3109309915C108010975 @default.
- W3109309915 hasConceptScore W3109309915C108583219 @default.
- W3109309915 hasConceptScore W3109309915C11413529 @default.
- W3109309915 hasConceptScore W3109309915C154945302 @default.
- W3109309915 hasConceptScore W3109309915C155032097 @default.
- W3109309915 hasConceptScore W3109309915C165464430 @default.
- W3109309915 hasConceptScore W3109309915C186370098 @default.
- W3109309915 hasConceptScore W3109309915C33923547 @default.
- W3109309915 hasConceptScore W3109309915C41008148 @default.
- W3109309915 hasConceptScore W3109309915C50644808 @default.
- W3109309915 hasConceptScore W3109309915C6557445 @default.
- W3109309915 hasConceptScore W3109309915C86803240 @default.
- W3109309915 hasFunder F4320322030 @default.
- W3109309915 hasFunder F4320322120 @default.
- W3109309915 hasLocation W31093099151 @default.
- W3109309915 hasOpenAccess W3109309915 @default.
- W3109309915 hasPrimaryLocation W31093099151 @default.
- W3109309915 hasRelatedWork W1540849649 @default.
- W3109309915 hasRelatedWork W2150503081 @default.
- W3109309915 hasRelatedWork W2187121994 @default.
- W3109309915 hasRelatedWork W2362189222 @default.
- W3109309915 hasRelatedWork W2391384657 @default.
- W3109309915 hasRelatedWork W2480330148 @default.
- W3109309915 hasRelatedWork W2804482613 @default.
- W3109309915 hasRelatedWork W3158487625 @default.
- W3109309915 hasRelatedWork W4210678178 @default.
- W3109309915 hasRelatedWork W2516580779 @default.
- W3109309915 hasVolume "4" @default.
- W3109309915 isParatext "false" @default.
- W3109309915 isRetracted "false" @default.
- W3109309915 magId "3109309915" @default.
- W3109309915 workType "article" @default.