Matches in SemOpenAlex for { <https://semopenalex.org/work/W3080135165> ?p ?o ?g. }
Showing items 1 to 97 of
97
with 100 items per page.
- W3080135165 abstract "Convolutional neural networks (CNNs) have achieved tremendous success in solving many challenging computer vision tasks. However, CNNs are extremely demanding for computation capability, memory space, and power capacity. This limits their usage to the cloud and prevents them from being deployed on edge devices with constrained resources and power. To tackle this problem, we propose craft distillation, a novel model compression approach that leverages both depthwise separable convolutions and knowledge distillation to significantly reduce the size of a highly complex model. Craft distillation has three advantages over existing model compression techniques. First, it does not require prior experiences on how to design a good “student model” for effective knowledge distillation. Second, it does not require specialized hardware support (e.g. ASIC or FPGA). Third, it is compatible with existing model compression techniques and can be used with pruning and quantization together to further reduce weight storage and arithmetic operations. Our experimental results show that with proper layer block replacement design and replacement strategy, craft distillation reduces the computational cost of VGG16 by 74.6% compared to the original dense models with negligible influence on accuracy." @default.
- W3080135165 created "2020-09-01" @default.
- W3080135165 creator A5002287514 @default.
- W3080135165 creator A5008451482 @default.
- W3080135165 creator A5048411184 @default.
- W3080135165 creator A5065047387 @default.
- W3080135165 date "2020-08-01" @default.
- W3080135165 modified "2023-09-26" @default.
- W3080135165 title "Craft Distillation: Layer-wise Convolutional Neural Network Distillation" @default.
- W3080135165 cites W1821462560 @default.
- W3080135165 cites W1972544340 @default.
- W3080135165 cites W2094756095 @default.
- W3080135165 cites W2114766824 @default.
- W3080135165 cites W2121775913 @default.
- W3080135165 cites W2125389748 @default.
- W3080135165 cites W2184188583 @default.
- W3080135165 cites W2194775991 @default.
- W3080135165 cites W2276486856 @default.
- W3080135165 cites W2403646140 @default.
- W3080135165 cites W2565851976 @default.
- W3080135165 cites W2585720638 @default.
- W3080135165 cites W2612445135 @default.
- W3080135165 cites W2612864759 @default.
- W3080135165 cites W2619096655 @default.
- W3080135165 cites W2625457103 @default.
- W3080135165 cites W2756085244 @default.
- W3080135165 cites W2766839578 @default.
- W3080135165 cites W2767864757 @default.
- W3080135165 cites W2769312802 @default.
- W3080135165 cites W2783873922 @default.
- W3080135165 cites W2787752464 @default.
- W3080135165 cites W2808373772 @default.
- W3080135165 cites W2950967261 @default.
- W3080135165 cites W2962965870 @default.
- W3080135165 cites W2963225922 @default.
- W3080135165 cites W2963674932 @default.
- W3080135165 cites W2964299589 @default.
- W3080135165 cites W3097096317 @default.
- W3080135165 doi "https://doi.org/10.1109/cscloud-edgecom49738.2020.00051" @default.
- W3080135165 hasPublicationYear "2020" @default.
- W3080135165 type Work @default.
- W3080135165 sameAs 3080135165 @default.
- W3080135165 citedByCount "2" @default.
- W3080135165 countsByYear W30801351652020 @default.
- W3080135165 countsByYear W30801351652021 @default.
- W3080135165 crossrefType "proceedings-article" @default.
- W3080135165 hasAuthorship W3080135165A5002287514 @default.
- W3080135165 hasAuthorship W3080135165A5008451482 @default.
- W3080135165 hasAuthorship W3080135165A5048411184 @default.
- W3080135165 hasAuthorship W3080135165A5065047387 @default.
- W3080135165 hasConcept C113775141 @default.
- W3080135165 hasConcept C11413529 @default.
- W3080135165 hasConcept C119857082 @default.
- W3080135165 hasConcept C149635348 @default.
- W3080135165 hasConcept C154945302 @default.
- W3080135165 hasConcept C178790620 @default.
- W3080135165 hasConcept C185592680 @default.
- W3080135165 hasConcept C204030448 @default.
- W3080135165 hasConcept C28855332 @default.
- W3080135165 hasConcept C41008148 @default.
- W3080135165 hasConcept C42935608 @default.
- W3080135165 hasConcept C45374587 @default.
- W3080135165 hasConcept C50644808 @default.
- W3080135165 hasConcept C77390884 @default.
- W3080135165 hasConcept C81363708 @default.
- W3080135165 hasConceptScore W3080135165C113775141 @default.
- W3080135165 hasConceptScore W3080135165C11413529 @default.
- W3080135165 hasConceptScore W3080135165C119857082 @default.
- W3080135165 hasConceptScore W3080135165C149635348 @default.
- W3080135165 hasConceptScore W3080135165C154945302 @default.
- W3080135165 hasConceptScore W3080135165C178790620 @default.
- W3080135165 hasConceptScore W3080135165C185592680 @default.
- W3080135165 hasConceptScore W3080135165C204030448 @default.
- W3080135165 hasConceptScore W3080135165C28855332 @default.
- W3080135165 hasConceptScore W3080135165C41008148 @default.
- W3080135165 hasConceptScore W3080135165C42935608 @default.
- W3080135165 hasConceptScore W3080135165C45374587 @default.
- W3080135165 hasConceptScore W3080135165C50644808 @default.
- W3080135165 hasConceptScore W3080135165C77390884 @default.
- W3080135165 hasConceptScore W3080135165C81363708 @default.
- W3080135165 hasLocation W30801351651 @default.
- W3080135165 hasOpenAccess W3080135165 @default.
- W3080135165 hasPrimaryLocation W30801351651 @default.
- W3080135165 hasRelatedWork W1995921138 @default.
- W3080135165 hasRelatedWork W2096844293 @default.
- W3080135165 hasRelatedWork W2134422574 @default.
- W3080135165 hasRelatedWork W2376218453 @default.
- W3080135165 hasRelatedWork W2786216825 @default.
- W3080135165 hasRelatedWork W2892054964 @default.
- W3080135165 hasRelatedWork W2926371525 @default.
- W3080135165 hasRelatedWork W3009083538 @default.
- W3080135165 hasRelatedWork W3010492628 @default.
- W3080135165 hasRelatedWork W4288375792 @default.
- W3080135165 isParatext "false" @default.
- W3080135165 isRetracted "false" @default.
- W3080135165 magId "3080135165" @default.
- W3080135165 workType "article" @default.