FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W3048668727> ?p ?o ?g. }

• W3048668727 abstract "Accelerating deep model training and inference is crucial in practice. Existing deep learning frameworks usually concentrate on optimizing training speed and pay fewer attentions to inference-specific optimizations. Actually, model inference differs from training in terms of computation, e.g. parameters are refreshed each gradient update step during training, but kept invariant during inference. These special characteristics of model inference open new opportunities for its optimization. In this paper, we propose a hardware-aware optimization framework, namely Woodpecker-DL (WPK), to accelerate inference by taking advantage of multiple joint optimizations from the perspectives of graph optimization, automated searches, domain-specific language (DSL) compiler techniques and system-level exploration. In WPK, we investigated two new automated search approaches based on genetic algorithm and reinforcement learning, respectively, to hunt the best operator code configurations targeting specific hardware. A customized DSL compiler is further attached to these search algorithms to generate efficient codes. To create an optimized inference plan, WPK systematically explores high-speed operator implementations from third-party libraries besides our automatically generated codes and singles out the best implementation per operator for use. Extensive experiments demonstrated that on a Tesla P100 GPU, we can achieve the maximum speedup of 5.40 over cuDNN and 1.63 over TVM on individual convolution operators, and run up to 1.18 times faster than TensorRT for end-to-end model inference." @default.
• W3048668727 created "2020-08-18" @default.
• W3048668727 creator A5009074414 @default.
• W3048668727 creator A5024522346 @default.
• W3048668727 creator A5066660227 @default.
• W3048668727 creator A5067942871 @default.
• W3048668727 creator A5070628952 @default.
• W3048668727 creator A5080577523 @default.
• W3048668727 creator A5085462851 @default.
• W3048668727 date "2020-08-11" @default.
• W3048668727 modified "2023-09-23" @default.
• W3048668727 title "Woodpecker-DL: Accelerating Deep Neural Networks via Hardware-Aware Multifaceted Optimizations." @default.
• W3048668727 cites W1659842140 @default.
• W3048668727 cites W1667652561 @default.
• W3048668727 cites W2055312318 @default.
• W3048668727 cites W2155503253 @default.
• W3048668727 cites W2155893237 @default.
• W3048668727 cites W2163605009 @default.
• W3048668727 cites W2186615578 @default.
• W3048668727 cites W2194775991 @default.
• W3048668727 cites W2402144811 @default.
• W3048668727 cites W2521727659 @default.
• W3048668727 cites W2613718673 @default.
• W3048668727 cites W2736601468 @default.
• W3048668727 cites W2767935072 @default.
• W3048668727 cites W2784561332 @default.
• W3048668727 cites W2786320458 @default.
• W3048668727 cites W2786414509 @default.
• W3048668727 cites W2798341898 @default.
• W3048668727 cites W2804032941 @default.
• W3048668727 cites W2806891462 @default.
• W3048668727 cites W2811374584 @default.
• W3048668727 cites W2897925395 @default.
• W3048668727 cites W2905135312 @default.
• W3048668727 cites W2954698171 @default.
• W3048668727 cites W2963094322 @default.
• W3048668727 cites W2963150697 @default.
• W3048668727 cites W2963947383 @default.
• W3048668727 cites W2963960923 @default.
• W3048668727 cites W2964043796 @default.
• W3048668727 cites W2969388332 @default.
• W3048668727 cites W2970971581 @default.
• W3048668727 cites W2981758446 @default.
• W3048668727 cites W2990152177 @default.
• W3048668727 hasPublicationYear "2020" @default.
• W3048668727 type Work @default.
• W3048668727 sameAs 3048668727 @default.
• W3048668727 citedByCount "0" @default.
• W3048668727 crossrefType "posted-content" @default.
• W3048668727 hasAuthorship W3048668727A5009074414 @default.
• W3048668727 hasAuthorship W3048668727A5024522346 @default.
• W3048668727 hasAuthorship W3048668727A5066660227 @default.
• W3048668727 hasAuthorship W3048668727A5067942871 @default.
• W3048668727 hasAuthorship W3048668727A5070628952 @default.
• W3048668727 hasAuthorship W3048668727A5080577523 @default.
• W3048668727 hasAuthorship W3048668727A5085462851 @default.
• W3048668727 hasConcept C108583219 @default.
• W3048668727 hasConcept C113775141 @default.
• W3048668727 hasConcept C119857082 @default.
• W3048668727 hasConcept C133162039 @default.
• W3048668727 hasConcept C154945302 @default.
• W3048668727 hasConcept C169590947 @default.
• W3048668727 hasConcept C173608175 @default.
• W3048668727 hasConcept C199360897 @default.
• W3048668727 hasConcept C26517878 @default.
• W3048668727 hasConcept C2776214188 @default.
• W3048668727 hasConcept C38652104 @default.
• W3048668727 hasConcept C41008148 @default.
• W3048668727 hasConcept C68339613 @default.
• W3048668727 hasConcept C80444323 @default.
• W3048668727 hasConceptScore W3048668727C108583219 @default.
• W3048668727 hasConceptScore W3048668727C113775141 @default.
• W3048668727 hasConceptScore W3048668727C119857082 @default.
• W3048668727 hasConceptScore W3048668727C133162039 @default.
• W3048668727 hasConceptScore W3048668727C154945302 @default.
• W3048668727 hasConceptScore W3048668727C169590947 @default.
• W3048668727 hasConceptScore W3048668727C173608175 @default.
• W3048668727 hasConceptScore W3048668727C199360897 @default.
• W3048668727 hasConceptScore W3048668727C26517878 @default.
• W3048668727 hasConceptScore W3048668727C2776214188 @default.
• W3048668727 hasConceptScore W3048668727C38652104 @default.
• W3048668727 hasConceptScore W3048668727C41008148 @default.
• W3048668727 hasConceptScore W3048668727C68339613 @default.
• W3048668727 hasConceptScore W3048668727C80444323 @default.
• W3048668727 hasLocation W30486687271 @default.
• W3048668727 hasOpenAccess W3048668727 @default.
• W3048668727 hasPrimaryLocation W30486687271 @default.
• W3048668727 hasRelatedWork W1988659054 @default.
• W3048668727 hasRelatedWork W2063774307 @default.
• W3048668727 hasRelatedWork W2314944927 @default.
• W3048668727 hasRelatedWork W2407123212 @default.
• W3048668727 hasRelatedWork W2553918877 @default.
• W3048668727 hasRelatedWork W2586408124 @default.
• W3048668727 hasRelatedWork W2606167063 @default.
• W3048668727 hasRelatedWork W2625265050 @default.
• W3048668727 hasRelatedWork W2950449148 @default.
• W3048668727 hasRelatedWork W2975489744 @default.
• W3048668727 hasRelatedWork W2987932087 @default.
• W3048668727 hasRelatedWork W2992034195 @default.
• W3048668727 hasRelatedWork W2995723163 @default.

• next