Matches in SemOpenAlex for { <https://semopenalex.org/work/W4321276845> ?p ?o ?g. }
Showing items 1 to 67 of
67
with 100 items per page.
- W4321276845 abstract "Parameter-efficient transfer learning (PETL) is an emerging research spot aimed at inexpensively adapting large-scale pre-trained models to downstream tasks. Recent advances have achieved great success in saving storage costs for various pre-trained models by updating a small number of parameters instead of full tuning. However, we notice that most existing PETL methods still incur non-negligible latency during inference. In this paper, we propose a parameter-efficient and computational friendly adapter for giant vision models, called RepAdapter. Specifically, we first prove that common adaptation modules can also be seamlessly integrated into most giant vision models via our structural re-parameterization, thereby achieving zero-cost during inference. We then investigate the sparse design and effective placement of adapter structure, helping our RepAdaper obtain other advantages in terms of parameter efficiency and performance. To validate RepAdapter, we conduct extensive experiments on 27 benchmark datasets of three vision tasks, i.e., image and video classifications and semantic segmentation. Experimental results show the superior performance and efficiency of RepAdapter than the state-of-the-art PETL methods. For instance, RepAdapter outperforms full tuning by +7.2% on average and saves up to 25% training time, 20% GPU memory, and 94.6% storage cost of ViT-B/16 on VTAB-1k. The generalization ability of RepAdapter is also well validated by a bunch of vision models. Our source code is released at https://github.com/luogen1996/RepAdapter." @default.
- W4321276845 created "2023-02-18" @default.
- W4321276845 creator A5010520524 @default.
- W4321276845 creator A5016080094 @default.
- W4321276845 creator A5020568230 @default.
- W4321276845 creator A5024327669 @default.
- W4321276845 creator A5029317372 @default.
- W4321276845 creator A5044689489 @default.
- W4321276845 creator A5046378812 @default.
- W4321276845 date "2023-02-16" @default.
- W4321276845 modified "2023-09-23" @default.
- W4321276845 title "Towards Efficient Visual Adaption via Structural Re-parameterization" @default.
- W4321276845 doi "https://doi.org/10.48550/arxiv.2302.08106" @default.
- W4321276845 hasPublicationYear "2023" @default.
- W4321276845 type Work @default.
- W4321276845 citedByCount "0" @default.
- W4321276845 crossrefType "posted-content" @default.
- W4321276845 hasAuthorship W4321276845A5010520524 @default.
- W4321276845 hasAuthorship W4321276845A5016080094 @default.
- W4321276845 hasAuthorship W4321276845A5020568230 @default.
- W4321276845 hasAuthorship W4321276845A5024327669 @default.
- W4321276845 hasAuthorship W4321276845A5029317372 @default.
- W4321276845 hasAuthorship W4321276845A5044689489 @default.
- W4321276845 hasAuthorship W4321276845A5046378812 @default.
- W4321276845 hasBestOaLocation W43212768451 @default.
- W4321276845 hasConcept C113775141 @default.
- W4321276845 hasConcept C119857082 @default.
- W4321276845 hasConcept C13280743 @default.
- W4321276845 hasConcept C154945302 @default.
- W4321276845 hasConcept C177284502 @default.
- W4321276845 hasConcept C185798385 @default.
- W4321276845 hasConcept C205649164 @default.
- W4321276845 hasConcept C2776214188 @default.
- W4321276845 hasConcept C41008148 @default.
- W4321276845 hasConcept C48044578 @default.
- W4321276845 hasConcept C77088390 @default.
- W4321276845 hasConcept C89600930 @default.
- W4321276845 hasConcept C9390403 @default.
- W4321276845 hasConceptScore W4321276845C113775141 @default.
- W4321276845 hasConceptScore W4321276845C119857082 @default.
- W4321276845 hasConceptScore W4321276845C13280743 @default.
- W4321276845 hasConceptScore W4321276845C154945302 @default.
- W4321276845 hasConceptScore W4321276845C177284502 @default.
- W4321276845 hasConceptScore W4321276845C185798385 @default.
- W4321276845 hasConceptScore W4321276845C205649164 @default.
- W4321276845 hasConceptScore W4321276845C2776214188 @default.
- W4321276845 hasConceptScore W4321276845C41008148 @default.
- W4321276845 hasConceptScore W4321276845C48044578 @default.
- W4321276845 hasConceptScore W4321276845C77088390 @default.
- W4321276845 hasConceptScore W4321276845C89600930 @default.
- W4321276845 hasConceptScore W4321276845C9390403 @default.
- W4321276845 hasLocation W43212768451 @default.
- W4321276845 hasOpenAccess W4321276845 @default.
- W4321276845 hasPrimaryLocation W43212768451 @default.
- W4321276845 hasRelatedWork W112744582 @default.
- W4321276845 hasRelatedWork W1992807924 @default.
- W4321276845 hasRelatedWork W2102111939 @default.
- W4321276845 hasRelatedWork W2511279186 @default.
- W4321276845 hasRelatedWork W2789601449 @default.
- W4321276845 hasRelatedWork W2960426010 @default.
- W4321276845 hasRelatedWork W3088526462 @default.
- W4321276845 hasRelatedWork W3214188003 @default.
- W4321276845 hasRelatedWork W3214520715 @default.
- W4321276845 hasRelatedWork W4321855229 @default.
- W4321276845 isParatext "false" @default.
- W4321276845 isRetracted "false" @default.
- W4321276845 workType "article" @default.