Matches in SemOpenAlex for { <https://semopenalex.org/work/W4312348045> ?p ?o ?g. }
- W4312348045 abstract "Image restoration tasks have witnessed great performance improvement in recent years by developing large deep models. Despite the outstanding performance, the heavy computation demanded by the deep models has restricted the application of image restoration. To lift the restriction, it is required to reduce the size of the networks while maintaining accuracy. Recently, N:M structured pruning has appeared as one of the effective and practical pruning approaches for making the model efficient with the accuracy constraint. However, it fails to account for different computational complexities and performance requirements for different layers of an image restoration network. To further optimize the trade-off between the efficiency and the restoration accuracy, we propose a novel pruning method that determines the pruning ratio for N:M structured sparsity at each layer. Extensive experimental results on super-resolution and deblurring tasks demonstrate the efficacy of our method which outperforms previous pruning methods significantly. PyTorch implementation for the proposed methods will be publicly available at https://github.com/JungHunOh/SLS_CVPR2022" @default.
- W4312348045 created "2023-01-04" @default.
- W4312348045 creator A5003484891 @default.
- W4312348045 creator A5003678844 @default.
- W4312348045 creator A5046504049 @default.
- W4312348045 creator A5047136540 @default.
- W4312348045 creator A5058394307 @default.
- W4312348045 creator A5073483751 @default.
- W4312348045 date "2022-06-01" @default.
- W4312348045 modified "2023-09-30" @default.
- W4312348045 title "Attentive Fine-Grained Structured Sparsity for Image Restoration" @default.
- W4312348045 cites W1791560514 @default.
- W4312348045 cites W1885185971 @default.
- W4312348045 cites W1930824406 @default.
- W4312348045 cites W2121927366 @default.
- W4312348045 cites W2560533888 @default.
- W4312348045 cites W2741137940 @default.
- W4312348045 cites W2798735168 @default.
- W4312348045 cites W2899244816 @default.
- W4312348045 cites W2924515500 @default.
- W4312348045 cites W2928560789 @default.
- W4312348045 cites W2954930822 @default.
- W4312348045 cites W2961218591 @default.
- W4312348045 cites W2962851801 @default.
- W4312348045 cites W2963163009 @default.
- W4312348045 cites W2963363373 @default.
- W4312348045 cites W2963372104 @default.
- W4312348045 cites W2963729050 @default.
- W4312348045 cites W2964030969 @default.
- W4312348045 cites W2964101377 @default.
- W4312348045 cites W2964233199 @default.
- W4312348045 cites W2964350391 @default.
- W4312348045 cites W2965217508 @default.
- W4312348045 cites W2965862774 @default.
- W4312348045 cites W2982795046 @default.
- W4312348045 cites W2984618279 @default.
- W4312348045 cites W2997884640 @default.
- W4312348045 cites W3034234149 @default.
- W4312348045 cites W3034513523 @default.
- W4312348045 cites W3034789174 @default.
- W4312348045 cites W3035377608 @default.
- W4312348045 cites W3035467254 @default.
- W4312348045 cites W3119817362 @default.
- W4312348045 cites W3133953507 @default.
- W4312348045 cites W3163767593 @default.
- W4312348045 cites W3170697543 @default.
- W4312348045 cites W3176997885 @default.
- W4312348045 cites W3178648163 @default.
- W4312348045 cites W4212804231 @default.
- W4312348045 doi "https://doi.org/10.1109/cvpr52688.2022.01715" @default.
- W4312348045 hasPublicationYear "2022" @default.
- W4312348045 type Work @default.
- W4312348045 citedByCount "3" @default.
- W4312348045 countsByYear W43123480452022 @default.
- W4312348045 countsByYear W43123480452023 @default.
- W4312348045 crossrefType "proceedings-article" @default.
- W4312348045 hasAuthorship W4312348045A5003484891 @default.
- W4312348045 hasAuthorship W4312348045A5003678844 @default.
- W4312348045 hasAuthorship W4312348045A5046504049 @default.
- W4312348045 hasAuthorship W4312348045A5047136540 @default.
- W4312348045 hasAuthorship W4312348045A5058394307 @default.
- W4312348045 hasAuthorship W4312348045A5073483751 @default.
- W4312348045 hasBestOaLocation W43123480452 @default.
- W4312348045 hasConcept C106430172 @default.
- W4312348045 hasConcept C108010975 @default.
- W4312348045 hasConcept C11413529 @default.
- W4312348045 hasConcept C115961682 @default.
- W4312348045 hasConcept C119857082 @default.
- W4312348045 hasConcept C154945302 @default.
- W4312348045 hasConcept C2524010 @default.
- W4312348045 hasConcept C2776036281 @default.
- W4312348045 hasConcept C2777693668 @default.
- W4312348045 hasConcept C33923547 @default.
- W4312348045 hasConcept C41008148 @default.
- W4312348045 hasConcept C45374587 @default.
- W4312348045 hasConcept C6557445 @default.
- W4312348045 hasConcept C86803240 @default.
- W4312348045 hasConcept C9417928 @default.
- W4312348045 hasConceptScore W4312348045C106430172 @default.
- W4312348045 hasConceptScore W4312348045C108010975 @default.
- W4312348045 hasConceptScore W4312348045C11413529 @default.
- W4312348045 hasConceptScore W4312348045C115961682 @default.
- W4312348045 hasConceptScore W4312348045C119857082 @default.
- W4312348045 hasConceptScore W4312348045C154945302 @default.
- W4312348045 hasConceptScore W4312348045C2524010 @default.
- W4312348045 hasConceptScore W4312348045C2776036281 @default.
- W4312348045 hasConceptScore W4312348045C2777693668 @default.
- W4312348045 hasConceptScore W4312348045C33923547 @default.
- W4312348045 hasConceptScore W4312348045C41008148 @default.
- W4312348045 hasConceptScore W4312348045C45374587 @default.
- W4312348045 hasConceptScore W4312348045C6557445 @default.
- W4312348045 hasConceptScore W4312348045C86803240 @default.
- W4312348045 hasConceptScore W4312348045C9417928 @default.
- W4312348045 hasLocation W43123480451 @default.
- W4312348045 hasLocation W43123480452 @default.
- W4312348045 hasOpenAccess W4312348045 @default.
- W4312348045 hasPrimaryLocation W43123480451 @default.
- W4312348045 hasRelatedWork W1978130285 @default.
- W4312348045 hasRelatedWork W1993928981 @default.
- W4312348045 hasRelatedWork W2008178540 @default.