Matches in SemOpenAlex for { <https://semopenalex.org/work/W2808253432> ?p ?o ?g. }
Showing items 1 to 76 of
76
with 100 items per page.
- W2808253432 abstract "Neural networks offer high-accuracy solutions to a range of problems, but are costly to run in production systems because of computational and memory requirements during a forward pass. Given a trained network, we propose a techique called Deep Learning Approximation to build a faster network in a tiny fraction of the time required for training by only manipulating the network structure and coefficients without requiring re-training or access to the training data. Speedup is achieved by by applying a sequential series of independent optimizations that reduce the floating-point operations (FLOPs) required to perform a forward pass. First, lossless optimizations are applied, followed by lossy approximations using singular value decomposition (SVD) and low-rank matrix decomposition. The optimal approximation is chosen by weighing the relative accuracy loss and FLOP reduction according to a single parameter specified by the user. On PASCAL VOC 2007 with the YOLO network, we show an end-to-end 2x speedup in a network forward pass with a 5% drop in mAP that can be re-gained by finetuning." @default.
- W2808253432 created "2018-06-21" @default.
- W2808253432 creator A5001687522 @default.
- W2808253432 date "2018-10-14" @default.
- W2808253432 modified "2023-09-27" @default.
- W2808253432 title "Deep Learning Approximation: Zero-Shot Neural Network Speedup" @default.
- W2808253432 cites W1686810756 @default.
- W2808253432 cites W2002555321 @default.
- W2808253432 cites W2117539524 @default.
- W2808253432 cites W2167215970 @default.
- W2808253432 cites W2194775991 @default.
- W2808253432 cites W2279098554 @default.
- W2808253432 cites W2513419314 @default.
- W2808253432 cites W2612445135 @default.
- W2808253432 cites W2766839578 @default.
- W2808253432 cites W2949117887 @default.
- W2808253432 cites W2950094539 @default.
- W2808253432 cites W2950967261 @default.
- W2808253432 cites W2951433694 @default.
- W2808253432 cites W2952197171 @default.
- W2808253432 cites W2963419583 @default.
- W2808253432 cites W2964228333 @default.
- W2808253432 cites W2964299589 @default.
- W2808253432 hasPublicationYear "2018" @default.
- W2808253432 type Work @default.
- W2808253432 sameAs 2808253432 @default.
- W2808253432 citedByCount "0" @default.
- W2808253432 crossrefType "posted-content" @default.
- W2808253432 hasAuthorship W2808253432A5001687522 @default.
- W2808253432 hasConcept C108583219 @default.
- W2808253432 hasConcept C11413529 @default.
- W2808253432 hasConcept C154945302 @default.
- W2808253432 hasConcept C173608175 @default.
- W2808253432 hasConcept C22789450 @default.
- W2808253432 hasConcept C3826847 @default.
- W2808253432 hasConcept C41008148 @default.
- W2808253432 hasConcept C50644808 @default.
- W2808253432 hasConcept C68339613 @default.
- W2808253432 hasConcept C84211073 @default.
- W2808253432 hasConceptScore W2808253432C108583219 @default.
- W2808253432 hasConceptScore W2808253432C11413529 @default.
- W2808253432 hasConceptScore W2808253432C154945302 @default.
- W2808253432 hasConceptScore W2808253432C173608175 @default.
- W2808253432 hasConceptScore W2808253432C22789450 @default.
- W2808253432 hasConceptScore W2808253432C3826847 @default.
- W2808253432 hasConceptScore W2808253432C41008148 @default.
- W2808253432 hasConceptScore W2808253432C50644808 @default.
- W2808253432 hasConceptScore W2808253432C68339613 @default.
- W2808253432 hasConceptScore W2808253432C84211073 @default.
- W2808253432 hasLocation W28082534321 @default.
- W2808253432 hasOpenAccess W2808253432 @default.
- W2808253432 hasPrimaryLocation W28082534321 @default.
- W2808253432 hasRelatedWork W1503452384 @default.
- W2808253432 hasRelatedWork W2891335318 @default.
- W2808253432 hasRelatedWork W2891491149 @default.
- W2808253432 hasRelatedWork W2892799969 @default.
- W2808253432 hasRelatedWork W2902991518 @default.
- W2808253432 hasRelatedWork W2923555562 @default.
- W2808253432 hasRelatedWork W2952634710 @default.
- W2808253432 hasRelatedWork W2962956675 @default.
- W2808253432 hasRelatedWork W2971667683 @default.
- W2808253432 hasRelatedWork W2973163720 @default.
- W2808253432 hasRelatedWork W2994782216 @default.
- W2808253432 hasRelatedWork W2998619828 @default.
- W2808253432 hasRelatedWork W3015220982 @default.
- W2808253432 hasRelatedWork W3081850639 @default.
- W2808253432 hasRelatedWork W3099365437 @default.
- W2808253432 hasRelatedWork W3123666279 @default.
- W2808253432 hasRelatedWork W3132407408 @default.
- W2808253432 hasRelatedWork W3137175096 @default.
- W2808253432 hasRelatedWork W3188045955 @default.
- W2808253432 hasRelatedWork W3204425029 @default.
- W2808253432 isParatext "false" @default.
- W2808253432 isRetracted "false" @default.
- W2808253432 magId "2808253432" @default.
- W2808253432 workType "article" @default.