FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W2783292547> ?p ?o ?g. }

• W2783292547 abstract "Going deeper and wider in neural architectures improves the accuracy, while the limited GPU DRAM places an undesired restriction on the network design domain. Deep Learning (DL) practitioners either need change to less desired network architectures, or nontrivially dissect a network across multiGPUs. These distract DL practitioners from concentrating on their original machine learning tasks. We present SuperNeurons: a dynamic GPU memory scheduling runtime to enable the network training far beyond the GPU DRAM capacity. SuperNeurons features 3 memory optimizations, textit{Liveness Analysis}, textit{Unified Tensor Pool}, and textit{Cost-Aware Recomputation}, all together they effectively reduce the network-wide peak memory usage down to the maximal memory usage among layers. We also address the performance issues in those memory saving techniques. Given the limited GPU DRAM, SuperNeurons not only provisions the necessary memory for the training, but also dynamically allocates the memory for convolution workspaces to achieve the high performance. Evaluations against Caffe, Torch, MXNet and TensorFlow have demonstrated that SuperNeurons trains at least 3.2432 deeper network than current ones with the leading performance. Particularly, SuperNeurons can train ResNet2500 that has $10^4$ basic network layers on a 12GB K40c." @default.
• W2783292547 created "2018-01-26" @default.
• W2783292547 creator A5010181097 @default.
• W2783292547 creator A5027254893 @default.
• W2783292547 creator A5034086130 @default.
• W2783292547 creator A5043209884 @default.
• W2783292547 creator A5051227924 @default.
• W2783292547 creator A5059605240 @default.
• W2783292547 creator A5062753948 @default.
• W2783292547 creator A5084877885 @default.
• W2783292547 date "2018-01-13" @default.
• W2783292547 modified "2023-09-23" @default.
• W2783292547 title "SuperNeurons: Dynamic GPU Memory Management for Training Deep Neural Networks" @default.
• W2783292547 cites W1548328233 @default.
• W2783292547 cites W1598984030 @default.
• W2783292547 cites W1667652561 @default.
• W2783292547 cites W1686810756 @default.
• W2783292547 cites W2097117768 @default.
• W2783292547 cites W2107878631 @default.
• W2783292547 cites W2112796928 @default.
• W2783292547 cites W2125389748 @default.
• W2783292547 cites W2155893237 @default.
• W2783292547 cites W2162390675 @default.
• W2783292547 cites W2163605009 @default.
• W2783292547 cites W2178615544 @default.
• W2783292547 cites W2186615578 @default.
• W2783292547 cites W2194775991 @default.
• W2783292547 cites W2208150243 @default.
• W2783292547 cites W2274287116 @default.
• W2783292547 cites W2330958039 @default.
• W2783292547 cites W2338908902 @default.
• W2783292547 cites W2737788279 @default.
• W2783292547 cites W2950656546 @default.
• W2783292547 cites W2951219045 @default.
• W2783292547 cites W2953384591 @default.
• W2783292547 cites W2963446712 @default.
• W2783292547 cites W2964174152 @default.
• W2783292547 cites W587794757 @default.
• W2783292547 cites W2298503502 @default.
• W2783292547 hasPublicationYear "2018" @default.
• W2783292547 type Work @default.
• W2783292547 sameAs 2783292547 @default.
• W2783292547 citedByCount "4" @default.
• W2783292547 countsByYear W27832925472019 @default.
• W2783292547 countsByYear W27832925472020 @default.
• W2783292547 crossrefType "proceedings-article" @default.
• W2783292547 hasAuthorship W2783292547A5010181097 @default.
• W2783292547 hasAuthorship W2783292547A5027254893 @default.
• W2783292547 hasAuthorship W2783292547A5034086130 @default.
• W2783292547 hasAuthorship W2783292547A5043209884 @default.
• W2783292547 hasAuthorship W2783292547A5051227924 @default.
• W2783292547 hasAuthorship W2783292547A5059605240 @default.
• W2783292547 hasAuthorship W2783292547A5062753948 @default.
• W2783292547 hasAuthorship W2783292547A5084877885 @default.
• W2783292547 hasConcept C108583219 @default.
• W2783292547 hasConcept C111919701 @default.
• W2783292547 hasConcept C118524514 @default.
• W2783292547 hasConcept C118702147 @default.
• W2783292547 hasConcept C120314980 @default.
• W2783292547 hasConcept C136085584 @default.
• W2783292547 hasConcept C154945302 @default.
• W2783292547 hasConcept C15569618 @default.
• W2783292547 hasConcept C162324750 @default.
• W2783292547 hasConcept C173608175 @default.
• W2783292547 hasConcept C176649486 @default.
• W2783292547 hasConcept C206729178 @default.
• W2783292547 hasConcept C21547014 @default.
• W2783292547 hasConcept C41008148 @default.
• W2783292547 hasConcept C50644808 @default.
• W2783292547 hasConcept C7366592 @default.
• W2783292547 hasConcept C74912251 @default.
• W2783292547 hasConcept C9390403 @default.
• W2783292547 hasConcept C98986596 @default.
• W2783292547 hasConceptScore W2783292547C108583219 @default.
• W2783292547 hasConceptScore W2783292547C111919701 @default.
• W2783292547 hasConceptScore W2783292547C118524514 @default.
• W2783292547 hasConceptScore W2783292547C118702147 @default.
• W2783292547 hasConceptScore W2783292547C120314980 @default.
• W2783292547 hasConceptScore W2783292547C136085584 @default.
• W2783292547 hasConceptScore W2783292547C154945302 @default.
• W2783292547 hasConceptScore W2783292547C15569618 @default.
• W2783292547 hasConceptScore W2783292547C162324750 @default.
• W2783292547 hasConceptScore W2783292547C173608175 @default.
• W2783292547 hasConceptScore W2783292547C176649486 @default.
• W2783292547 hasConceptScore W2783292547C206729178 @default.
• W2783292547 hasConceptScore W2783292547C21547014 @default.
• W2783292547 hasConceptScore W2783292547C41008148 @default.
• W2783292547 hasConceptScore W2783292547C50644808 @default.
• W2783292547 hasConceptScore W2783292547C7366592 @default.
• W2783292547 hasConceptScore W2783292547C74912251 @default.
• W2783292547 hasConceptScore W2783292547C9390403 @default.
• W2783292547 hasConceptScore W2783292547C98986596 @default.
• W2783292547 hasOpenAccess W2783292547 @default.
• W2783292547 hasRelatedWork W1534731610 @default.
• W2783292547 hasRelatedWork W1974307224 @default.
• W2783292547 hasRelatedWork W1979790909 @default.
• W2783292547 hasRelatedWork W2021009496 @default.
• W2783292547 hasRelatedWork W2052537642 @default.
• W2783292547 hasRelatedWork W2163605009 @default.
• W2783292547 hasRelatedWork W2275997533 @default.

• next