Matches in SemOpenAlex for { <https://semopenalex.org/work/W4312717168> ?p ?o ?g. }
Showing items 1 to 89 of
89
with 100 items per page.
- W4312717168 endingPage "327" @default.
- W4312717168 startingPage "312" @default.
- W4312717168 abstract "Recently low-precision deep learning accelerators (DLAs) have become popular due to their advantages in chip area and energy consumption, yet the low-precision quantized models on these DLAs bring in severe accuracy degradation. One way to achieve both high accuracy and efficient inference is to deploy high-precision neural networks on low-precision DLAs, which is rarely studied. In this paper, we propose the PArallel Low-precision Quantization (PalQuant) method that approximates high-precision computations via learning parallel low-precision representations from scratch. In addition, we present a novel cyclic shuffle module to boost the cross-group information communication between parallel low-precision groups. Extensive experiments demonstrate that PalQuant has superior performance to state-of-the-art quantization methods in both accuracy and inference speed, e.g., for ResNet-18 network quantization, PalQuant can obtain 0.52% higher accuracy and 1.78 $$times $$ speedup simultaneously over their 4-bit counter-part on a state-of-the-art 2-bit accelerator. Code is available at https://github.com/huqinghao/PalQuant ." @default.
- W4312717168 created "2023-01-05" @default.
- W4312717168 creator A5000727470 @default.
- W4312717168 creator A5034605242 @default.
- W4312717168 creator A5065752878 @default.
- W4312717168 creator A5090113653 @default.
- W4312717168 date "2022-01-01" @default.
- W4312717168 modified "2023-09-26" @default.
- W4312717168 title "PalQuant: Accelerating High-Precision Networks on Low-Precision Accelerators" @default.
- W4312717168 cites W2094756095 @default.
- W4312717168 cites W2194775991 @default.
- W4312717168 cites W2276486856 @default.
- W4312717168 cites W2289252105 @default.
- W4312717168 cites W2300242332 @default.
- W4312717168 cites W2585560244 @default.
- W4312717168 cites W2586654419 @default.
- W4312717168 cites W2612864759 @default.
- W4312717168 cites W2845210056 @default.
- W4312717168 cites W2884150179 @default.
- W4312717168 cites W2931118404 @default.
- W4312717168 cites W2946572707 @default.
- W4312717168 cites W2963145956 @default.
- W4312717168 cites W2963367920 @default.
- W4312717168 cites W2963480671 @default.
- W4312717168 cites W2981774126 @default.
- W4312717168 cites W2987956272 @default.
- W4312717168 cites W3004061291 @default.
- W4312717168 cites W3091813304 @default.
- W4312717168 cites W3102169921 @default.
- W4312717168 cites W3173567919 @default.
- W4312717168 cites W3176946833 @default.
- W4312717168 cites W4253012315 @default.
- W4312717168 doi "https://doi.org/10.1007/978-3-031-20083-0_19" @default.
- W4312717168 hasPublicationYear "2022" @default.
- W4312717168 type Work @default.
- W4312717168 citedByCount "0" @default.
- W4312717168 crossrefType "book-chapter" @default.
- W4312717168 hasAuthorship W4312717168A5000727470 @default.
- W4312717168 hasAuthorship W4312717168A5034605242 @default.
- W4312717168 hasAuthorship W4312717168A5065752878 @default.
- W4312717168 hasAuthorship W4312717168A5090113653 @default.
- W4312717168 hasBestOaLocation W43127171682 @default.
- W4312717168 hasConcept C108583219 @default.
- W4312717168 hasConcept C113775141 @default.
- W4312717168 hasConcept C11413529 @default.
- W4312717168 hasConcept C133095886 @default.
- W4312717168 hasConcept C154945302 @default.
- W4312717168 hasConcept C173608175 @default.
- W4312717168 hasConcept C177264268 @default.
- W4312717168 hasConcept C199360897 @default.
- W4312717168 hasConcept C2776214188 @default.
- W4312717168 hasConcept C2776760102 @default.
- W4312717168 hasConcept C28855332 @default.
- W4312717168 hasConcept C41008148 @default.
- W4312717168 hasConcept C45374587 @default.
- W4312717168 hasConcept C68339613 @default.
- W4312717168 hasConceptScore W4312717168C108583219 @default.
- W4312717168 hasConceptScore W4312717168C113775141 @default.
- W4312717168 hasConceptScore W4312717168C11413529 @default.
- W4312717168 hasConceptScore W4312717168C133095886 @default.
- W4312717168 hasConceptScore W4312717168C154945302 @default.
- W4312717168 hasConceptScore W4312717168C173608175 @default.
- W4312717168 hasConceptScore W4312717168C177264268 @default.
- W4312717168 hasConceptScore W4312717168C199360897 @default.
- W4312717168 hasConceptScore W4312717168C2776214188 @default.
- W4312717168 hasConceptScore W4312717168C2776760102 @default.
- W4312717168 hasConceptScore W4312717168C28855332 @default.
- W4312717168 hasConceptScore W4312717168C41008148 @default.
- W4312717168 hasConceptScore W4312717168C45374587 @default.
- W4312717168 hasConceptScore W4312717168C68339613 @default.
- W4312717168 hasLocation W43127171681 @default.
- W4312717168 hasLocation W43127171682 @default.
- W4312717168 hasOpenAccess W4312717168 @default.
- W4312717168 hasPrimaryLocation W43127171681 @default.
- W4312717168 hasRelatedWork W1509211761 @default.
- W4312717168 hasRelatedWork W156843270 @default.
- W4312717168 hasRelatedWork W1905659066 @default.
- W4312717168 hasRelatedWork W2007449167 @default.
- W4312717168 hasRelatedWork W3196579076 @default.
- W4312717168 hasRelatedWork W3216454443 @default.
- W4312717168 hasRelatedWork W4296563896 @default.
- W4312717168 hasRelatedWork W4301264193 @default.
- W4312717168 hasRelatedWork W4312717168 @default.
- W4312717168 hasRelatedWork W4322616423 @default.
- W4312717168 isParatext "false" @default.
- W4312717168 isRetracted "false" @default.
- W4312717168 workType "book-chapter" @default.