Matches in SemOpenAlex for { <https://semopenalex.org/work/W2786406308> ?p ?o ?g. }
Showing items 1 to 90 of
90
with 100 items per page.
- W2786406308 abstract "There are many applications scenarios for which the computational performance and memory footprint of the prediction phase of Deep Neural Networks (DNNs) need to be optimized. Binary Deep Neural Networks (BDNNs) have been shown to be an effective way of achieving this objective. In this paper, we show how Convolutional Neural Networks (CNNs) can be implemented using binary representations. Espresso is a compact, yet powerful library written in C/CUDA that features all the functionalities required for the forward propagation of CNNs, in a binary file less than 400KB, without any external dependencies. Although it is mainly designed to take advantage of massive GPU parallelism, Espresso also provides an equivalent CPU implementation for CNNs. Espresso provides special convolutional and dense layers for BCNNs, leveraging bit-packing and bit-wise computations for efficient execution. These techniques provide a speed-up of matrix-multiplication routines, and at the same time, reduce memory usage when storing parameters and activations. We experimentally show that Espresso is significantly faster than existing implementations of optimized binary neural networks (~ 2 orders of magnitude). Espresso is released under the Apache 2.0 license and is available at http://github.com/organization/project." @default.
- W2786406308 created "2018-02-23" @default.
- W2786406308 creator A5010445045 @default.
- W2786406308 creator A5037094498 @default.
- W2786406308 creator A5048007228 @default.
- W2786406308 date "2018-02-15" @default.
- W2786406308 modified "2023-09-23" @default.
- W2786406308 title "Espresso: Efficient Forward Propagation for Binary Deep Neural Networks" @default.
- W2786406308 cites W2002257715 @default.
- W2786406308 cites W2099021415 @default.
- W2786406308 cites W2108598243 @default.
- W2786406308 cites W2160815625 @default.
- W2786406308 cites W2194775991 @default.
- W2786406308 cites W2279098554 @default.
- W2786406308 cites W2285660444 @default.
- W2786406308 cites W2335728318 @default.
- W2786406308 cites W2499931820 @default.
- W2786406308 cites W2963114950 @default.
- W2786406308 cites W2964299589 @default.
- W2786406308 cites W2964308564 @default.
- W2786406308 hasPublicationYear "2018" @default.
- W2786406308 type Work @default.
- W2786406308 sameAs 2786406308 @default.
- W2786406308 citedByCount "6" @default.
- W2786406308 countsByYear W27864063082019 @default.
- W2786406308 countsByYear W27864063082020 @default.
- W2786406308 countsByYear W27864063082021 @default.
- W2786406308 crossrefType "proceedings-article" @default.
- W2786406308 hasAuthorship W2786406308A5010445045 @default.
- W2786406308 hasAuthorship W2786406308A5037094498 @default.
- W2786406308 hasAuthorship W2786406308A5048007228 @default.
- W2786406308 hasConcept C108583219 @default.
- W2786406308 hasConcept C111919701 @default.
- W2786406308 hasConcept C121332964 @default.
- W2786406308 hasConcept C154945302 @default.
- W2786406308 hasConcept C17349429 @default.
- W2786406308 hasConcept C173608175 @default.
- W2786406308 hasConcept C2778119891 @default.
- W2786406308 hasConcept C33923547 @default.
- W2786406308 hasConcept C41008148 @default.
- W2786406308 hasConcept C48372109 @default.
- W2786406308 hasConcept C50644808 @default.
- W2786406308 hasConcept C62520636 @default.
- W2786406308 hasConcept C74912251 @default.
- W2786406308 hasConcept C81363708 @default.
- W2786406308 hasConcept C84114770 @default.
- W2786406308 hasConcept C94375191 @default.
- W2786406308 hasConceptScore W2786406308C108583219 @default.
- W2786406308 hasConceptScore W2786406308C111919701 @default.
- W2786406308 hasConceptScore W2786406308C121332964 @default.
- W2786406308 hasConceptScore W2786406308C154945302 @default.
- W2786406308 hasConceptScore W2786406308C17349429 @default.
- W2786406308 hasConceptScore W2786406308C173608175 @default.
- W2786406308 hasConceptScore W2786406308C2778119891 @default.
- W2786406308 hasConceptScore W2786406308C33923547 @default.
- W2786406308 hasConceptScore W2786406308C41008148 @default.
- W2786406308 hasConceptScore W2786406308C48372109 @default.
- W2786406308 hasConceptScore W2786406308C50644808 @default.
- W2786406308 hasConceptScore W2786406308C62520636 @default.
- W2786406308 hasConceptScore W2786406308C74912251 @default.
- W2786406308 hasConceptScore W2786406308C81363708 @default.
- W2786406308 hasConceptScore W2786406308C84114770 @default.
- W2786406308 hasConceptScore W2786406308C94375191 @default.
- W2786406308 hasLocation W27864063081 @default.
- W2786406308 hasOpenAccess W2786406308 @default.
- W2786406308 hasPrimaryLocation W27864063081 @default.
- W2786406308 hasRelatedWork W2194775991 @default.
- W2786406308 hasRelatedWork W2300242332 @default.
- W2786406308 hasRelatedWork W2319920447 @default.
- W2786406308 hasRelatedWork W2405920868 @default.
- W2786406308 hasRelatedWork W2469490737 @default.
- W2786406308 hasRelatedWork W2611289746 @default.
- W2786406308 hasRelatedWork W2618591429 @default.
- W2786406308 hasRelatedWork W2690141065 @default.
- W2786406308 hasRelatedWork W2773313889 @default.
- W2786406308 hasRelatedWork W2786771851 @default.
- W2786406308 hasRelatedWork W2963114950 @default.
- W2786406308 hasRelatedWork W2964008506 @default.
- W2786406308 hasRelatedWork W2969868335 @default.
- W2786406308 hasRelatedWork W3008796096 @default.
- W2786406308 hasRelatedWork W3045744481 @default.
- W2786406308 hasRelatedWork W3045877795 @default.
- W2786406308 hasRelatedWork W3115410382 @default.
- W2786406308 hasRelatedWork W3134055916 @default.
- W2786406308 hasRelatedWork W3151612890 @default.
- W2786406308 hasRelatedWork W3185074080 @default.
- W2786406308 isParatext "false" @default.
- W2786406308 isRetracted "false" @default.
- W2786406308 magId "2786406308" @default.
- W2786406308 workType "article" @default.