Matches in SemOpenAlex for { <https://semopenalex.org/work/W2912168260> ?p ?o ?g. }
Showing items 1 to 89 of
89
with 100 items per page.
- W2912168260 abstract "Convolutional neural networks require significant memory bandwidth and storage for intermediate computations, apart from substantial computing resources. Neural network quantization has significant benefits in reducing the amount of intermediate results, but it often requires the full datasets and time-consuming fine tuning to recover the accuracy lost after quantization. This paper introduces the first practical 4-bit post training quantization approach: it does not involve training the quantized model (fine-tuning), nor it requires the availability of the full dataset. We target the quantization of both activations and weights and suggest three complementary methods for minimizing quantization error at the tensor level, two of whom obtain a closed-form analytical solution. Combining these methods, our approach achieves accuracy that is just a few percents less the state-of-the-art baseline across a wide range of convolutional models. The source code to replicate all experiments is available on GitHub: url{this https URL}." @default.
- W2912168260 created "2019-02-21" @default.
- W2912168260 creator A5025123281 @default.
- W2912168260 creator A5030587898 @default.
- W2912168260 creator A5057489819 @default.
- W2912168260 creator A5091675091 @default.
- W2912168260 date "2018-10-02" @default.
- W2912168260 modified "2023-09-27" @default.
- W2912168260 title "Post-training 4-bit quantization of convolution networks for rapid-deployment" @default.
- W2912168260 cites W2043481157 @default.
- W2912168260 cites W2161160262 @default.
- W2912168260 cites W2161758346 @default.
- W2912168260 cites W2300242332 @default.
- W2912168260 cites W2469490737 @default.
- W2912168260 cites W2786771851 @default.
- W2912168260 cites W2798544842 @default.
- W2912168260 cites W2809624076 @default.
- W2912168260 cites W2897049076 @default.
- W2912168260 cites W2904908132 @default.
- W2912168260 cites W2946943679 @default.
- W2912168260 cites W2963122961 @default.
- W2912168260 cites W2964008506 @default.
- W2912168260 cites W2964164125 @default.
- W2912168260 cites W2964299589 @default.
- W2912168260 cites W2998218113 @default.
- W2912168260 hasPublicationYear "2018" @default.
- W2912168260 type Work @default.
- W2912168260 sameAs 2912168260 @default.
- W2912168260 citedByCount "42" @default.
- W2912168260 countsByYear W29121682602019 @default.
- W2912168260 countsByYear W29121682602020 @default.
- W2912168260 countsByYear W29121682602021 @default.
- W2912168260 countsByYear W29121682602022 @default.
- W2912168260 crossrefType "posted-content" @default.
- W2912168260 hasAuthorship W2912168260A5025123281 @default.
- W2912168260 hasAuthorship W2912168260A5030587898 @default.
- W2912168260 hasAuthorship W2912168260A5057489819 @default.
- W2912168260 hasAuthorship W2912168260A5091675091 @default.
- W2912168260 hasConcept C105339364 @default.
- W2912168260 hasConcept C105795698 @default.
- W2912168260 hasConcept C111919701 @default.
- W2912168260 hasConcept C113775141 @default.
- W2912168260 hasConcept C11413529 @default.
- W2912168260 hasConcept C154945302 @default.
- W2912168260 hasConcept C2781162219 @default.
- W2912168260 hasConcept C28855332 @default.
- W2912168260 hasConcept C33923547 @default.
- W2912168260 hasConcept C41008148 @default.
- W2912168260 hasConcept C45374587 @default.
- W2912168260 hasConcept C81363708 @default.
- W2912168260 hasConceptScore W2912168260C105339364 @default.
- W2912168260 hasConceptScore W2912168260C105795698 @default.
- W2912168260 hasConceptScore W2912168260C111919701 @default.
- W2912168260 hasConceptScore W2912168260C113775141 @default.
- W2912168260 hasConceptScore W2912168260C11413529 @default.
- W2912168260 hasConceptScore W2912168260C154945302 @default.
- W2912168260 hasConceptScore W2912168260C2781162219 @default.
- W2912168260 hasConceptScore W2912168260C28855332 @default.
- W2912168260 hasConceptScore W2912168260C33923547 @default.
- W2912168260 hasConceptScore W2912168260C41008148 @default.
- W2912168260 hasConceptScore W2912168260C45374587 @default.
- W2912168260 hasConceptScore W2912168260C81363708 @default.
- W2912168260 hasLocation W29121682601 @default.
- W2912168260 hasOpenAccess W2912168260 @default.
- W2912168260 hasPrimaryLocation W29121682601 @default.
- W2912168260 hasRelatedWork W1686810756 @default.
- W2912168260 hasRelatedWork W2108598243 @default.
- W2912168260 hasRelatedWork W2194775991 @default.
- W2912168260 hasRelatedWork W2242818861 @default.
- W2912168260 hasRelatedWork W2300242332 @default.
- W2912168260 hasRelatedWork W2469490737 @default.
- W2912168260 hasRelatedWork W2612445135 @default.
- W2912168260 hasRelatedWork W2786771851 @default.
- W2912168260 hasRelatedWork W2809624076 @default.
- W2912168260 hasRelatedWork W2884150179 @default.
- W2912168260 hasRelatedWork W2896718327 @default.
- W2912168260 hasRelatedWork W2963114950 @default.
- W2912168260 hasRelatedWork W2963163009 @default.
- W2912168260 hasRelatedWork W2963374099 @default.
- W2912168260 hasRelatedWork W2963480671 @default.
- W2912168260 hasRelatedWork W2964264300 @default.
- W2912168260 hasRelatedWork W2964299589 @default.
- W2912168260 hasRelatedWork W2981751377 @default.
- W2912168260 hasRelatedWork W2998218113 @default.
- W2912168260 hasRelatedWork W3034940165 @default.
- W2912168260 isParatext "false" @default.
- W2912168260 isRetracted "false" @default.
- W2912168260 magId "2912168260" @default.
- W2912168260 workType "article" @default.