Matches in SemOpenAlex for { <https://semopenalex.org/work/W2900493439> ?p ?o ?g. }
- W2900493439 abstract "We present a filter correlation based model compression approach for deep convolutional neural networks. Our approach iteratively identifies pairs of filters with the largest pairwise correlations and drops one of the filters from each such pair. However, instead of discarding one of the filters from each such pair naively, the model is re-optimized to make the filters in these pairs maximally correlated, so that discarding one of the filters from the pair results in minimal information loss. Moreover, after discarding the filters in each round, we further finetune the model to recover from the potential small loss incurred by the compression. We evaluate our proposed approach using a comprehensive set of experiments and ablation studies. Our compression method yields state-of-the-art FLOPs compression rates on various benchmarks, such as LeNet-5, VGG-16, and ResNet-50,56, while still achieving excellent predictive performance for tasks such as object detection on benchmark datasets." @default.
- W2900493439 created "2018-11-29" @default.
- W2900493439 creator A5007109424 @default.
- W2900493439 creator A5007774595 @default.
- W2900493439 creator A5029651089 @default.
- W2900493439 creator A5050600918 @default.
- W2900493439 date "2018-11-26" @default.
- W2900493439 modified "2023-09-26" @default.
- W2900493439 title "Leveraging Filter Correlations for Deep Model Compression" @default.
- W2900493439 cites W1861492603 @default.
- W2900493439 cites W1902041153 @default.
- W2900493439 cites W2002427601 @default.
- W2900493439 cites W2112796928 @default.
- W2900493439 cites W2114766824 @default.
- W2900493439 cites W2117539524 @default.
- W2900493439 cites W2125389748 @default.
- W2900493439 cites W2167215970 @default.
- W2900493439 cites W2172166488 @default.
- W2900493439 cites W2194775991 @default.
- W2900493439 cites W2300242332 @default.
- W2900493439 cites W2495425901 @default.
- W2900493439 cites W2515385951 @default.
- W2900493439 cites W2520760693 @default.
- W2900493439 cites W2554931888 @default.
- W2900493439 cites W2565639579 @default.
- W2900493439 cites W2572009792 @default.
- W2900493439 cites W2582745083 @default.
- W2900493439 cites W2613718673 @default.
- W2900493439 cites W2614986686 @default.
- W2900493439 cites W2619444510 @default.
- W2900493439 cites W2619890685 @default.
- W2900493439 cites W2625674597 @default.
- W2900493439 cites W2707890836 @default.
- W2900493439 cites W2757143157 @default.
- W2900493439 cites W2769182847 @default.
- W2900493439 cites W2769312802 @default.
- W2900493439 cites W2778053709 @default.
- W2900493439 cites W2787752464 @default.
- W2900493439 cites W2788653909 @default.
- W2900493439 cites W2788715907 @default.
- W2900493439 cites W2793035069 @default.
- W2900493439 cites W2799197246 @default.
- W2900493439 cites W2808168148 @default.
- W2900493439 cites W2883070812 @default.
- W2900493439 cites W2883111419 @default.
- W2900493439 cites W2886851211 @default.
- W2900493439 cites W2899408199 @default.
- W2900493439 cites W2921106662 @default.
- W2900493439 cites W2949273893 @default.
- W2900493439 cites W2950967261 @default.
- W2900493439 cites W2962684187 @default.
- W2900493439 cites W2962723643 @default.
- W2900493439 cites W2962835968 @default.
- W2900493439 cites W2962851801 @default.
- W2900493439 cites W2962944050 @default.
- W2900493439 cites W2962980542 @default.
- W2900493439 cites W2963000224 @default.
- W2900493439 cites W2963117513 @default.
- W2900493439 cites W2963118768 @default.
- W2900493439 cites W2963145730 @default.
- W2900493439 cites W2963363373 @default.
- W2900493439 cites W2963545832 @default.
- W2900493439 cites W2963559821 @default.
- W2900493439 cites W2964065888 @default.
- W2900493439 cites W2964078140 @default.
- W2900493439 cites W2964121937 @default.
- W2900493439 cites W2964233199 @default.
- W2900493439 cites W2964299589 @default.
- W2900493439 cites W2965955046 @default.
- W2900493439 cites W2981421796 @default.
- W2900493439 cites W2982536526 @default.
- W2900493439 cites W2983868233 @default.
- W2900493439 cites W3106250896 @default.
- W2900493439 cites W3013573966 @default.
- W2900493439 hasPublicationYear "2018" @default.
- W2900493439 type Work @default.
- W2900493439 sameAs 2900493439 @default.
- W2900493439 citedByCount "6" @default.
- W2900493439 countsByYear W29004934392019 @default.
- W2900493439 countsByYear W29004934392021 @default.
- W2900493439 crossrefType "posted-content" @default.
- W2900493439 hasAuthorship W2900493439A5007109424 @default.
- W2900493439 hasAuthorship W2900493439A5007774595 @default.
- W2900493439 hasAuthorship W2900493439A5029651089 @default.
- W2900493439 hasAuthorship W2900493439A5050600918 @default.
- W2900493439 hasConcept C106131492 @default.
- W2900493439 hasConcept C11413529 @default.
- W2900493439 hasConcept C127413603 @default.
- W2900493439 hasConcept C13280743 @default.
- W2900493439 hasConcept C153180895 @default.
- W2900493439 hasConcept C154945302 @default.
- W2900493439 hasConcept C159985019 @default.
- W2900493439 hasConcept C171146098 @default.
- W2900493439 hasConcept C173608175 @default.
- W2900493439 hasConcept C177264268 @default.
- W2900493439 hasConcept C180016635 @default.
- W2900493439 hasConcept C184898388 @default.
- W2900493439 hasConcept C185798385 @default.
- W2900493439 hasConcept C192562407 @default.
- W2900493439 hasConcept C199360897 @default.