Matches in SemOpenAlex for { <https://semopenalex.org/work/W4289463412> ?p ?o ?g. }
Showing items 1 to 71 of
71
with 100 items per page.
- W4289463412 abstract "Recent advances in pruning of neural networks have made it possible to remove a large number of filters or weights without any perceptible drop in accuracy. The number of parameters and that of FLOPs are usually the reported metrics to measure the quality of the pruned models. However, the gain in speed for these pruned models is often overlooked in the literature due to the complex nature of latency measurements. In this paper, we show the limitation of filter pruning methods in terms of latency reduction and propose LayerPrune framework. LayerPrune presents a set of layer pruning methods based on different criteria that achieve higher latency reduction than filter pruning methods on similar accuracy. The advantage of layer pruning over filter pruning in terms of latency reduction is a result of the fact that the former is not constrained by the original model's depth and thus allows for a larger range of latency reduction. For each filter pruning method we examined, we use the same filter importance criterion to calculate a per-layer importance score in one-shot. We then prune the least important layers and fine-tune the shallower model which obtains comparable or better accuracy than its filter-based pruning counterpart. This one-shot process allows to remove layers from single path networks like VGG before fine-tuning, unlike in iterative filter pruning, a minimum number of filters per layer is required to allow for data flow which constraint the search space. To the best of our knowledge, we are the first to examine the effect of pruning methods on latency metric instead of FLOPs for multiple networks, datasets and hardware targets. LayerPrune also outperforms handcrafted architectures such as Shufflenet, MobileNet, MNASNet and ResNet18 by 7.3%, 4.6%, 2.8% and 0.5% respectively on similar latency budget on ImageNet dataset." @default.
- W4289463412 created "2022-08-02" @default.
- W4289463412 creator A5030303076 @default.
- W4289463412 creator A5031013766 @default.
- W4289463412 creator A5069849278 @default.
- W4289463412 creator A5082800075 @default.
- W4289463412 creator A5091387172 @default.
- W4289463412 date "2020-07-10" @default.
- W4289463412 modified "2023-10-16" @default.
- W4289463412 title "To Filter Prune, or to Layer Prune, That Is The Question" @default.
- W4289463412 doi "https://doi.org/10.48550/arxiv.2007.05667" @default.
- W4289463412 hasPublicationYear "2020" @default.
- W4289463412 type Work @default.
- W4289463412 citedByCount "0" @default.
- W4289463412 crossrefType "posted-content" @default.
- W4289463412 hasAuthorship W4289463412A5030303076 @default.
- W4289463412 hasAuthorship W4289463412A5031013766 @default.
- W4289463412 hasAuthorship W4289463412A5069849278 @default.
- W4289463412 hasAuthorship W4289463412A5082800075 @default.
- W4289463412 hasAuthorship W4289463412A5091387172 @default.
- W4289463412 hasBestOaLocation W42894634121 @default.
- W4289463412 hasConcept C106131492 @default.
- W4289463412 hasConcept C108010975 @default.
- W4289463412 hasConcept C111335779 @default.
- W4289463412 hasConcept C11413529 @default.
- W4289463412 hasConcept C13107197 @default.
- W4289463412 hasConcept C154945302 @default.
- W4289463412 hasConcept C173608175 @default.
- W4289463412 hasConcept C22597639 @default.
- W4289463412 hasConcept C2524010 @default.
- W4289463412 hasConcept C31972630 @default.
- W4289463412 hasConcept C33923547 @default.
- W4289463412 hasConcept C3826847 @default.
- W4289463412 hasConcept C41008148 @default.
- W4289463412 hasConcept C6557445 @default.
- W4289463412 hasConcept C76155785 @default.
- W4289463412 hasConcept C82876162 @default.
- W4289463412 hasConcept C86803240 @default.
- W4289463412 hasConceptScore W4289463412C106131492 @default.
- W4289463412 hasConceptScore W4289463412C108010975 @default.
- W4289463412 hasConceptScore W4289463412C111335779 @default.
- W4289463412 hasConceptScore W4289463412C11413529 @default.
- W4289463412 hasConceptScore W4289463412C13107197 @default.
- W4289463412 hasConceptScore W4289463412C154945302 @default.
- W4289463412 hasConceptScore W4289463412C173608175 @default.
- W4289463412 hasConceptScore W4289463412C22597639 @default.
- W4289463412 hasConceptScore W4289463412C2524010 @default.
- W4289463412 hasConceptScore W4289463412C31972630 @default.
- W4289463412 hasConceptScore W4289463412C33923547 @default.
- W4289463412 hasConceptScore W4289463412C3826847 @default.
- W4289463412 hasConceptScore W4289463412C41008148 @default.
- W4289463412 hasConceptScore W4289463412C6557445 @default.
- W4289463412 hasConceptScore W4289463412C76155785 @default.
- W4289463412 hasConceptScore W4289463412C82876162 @default.
- W4289463412 hasConceptScore W4289463412C86803240 @default.
- W4289463412 hasLocation W42894634121 @default.
- W4289463412 hasOpenAccess W4289463412 @default.
- W4289463412 hasPrimaryLocation W42894634121 @default.
- W4289463412 hasRelatedWork W12246666 @default.
- W4289463412 hasRelatedWork W12413966 @default.
- W4289463412 hasRelatedWork W1419050 @default.
- W4289463412 hasRelatedWork W1674447 @default.
- W4289463412 hasRelatedWork W325970 @default.
- W4289463412 hasRelatedWork W6027241 @default.
- W4289463412 hasRelatedWork W8226277 @default.
- W4289463412 hasRelatedWork W8394581 @default.
- W4289463412 hasRelatedWork W8451425 @default.
- W4289463412 hasRelatedWork W8892619 @default.
- W4289463412 isParatext "false" @default.
- W4289463412 isRetracted "false" @default.
- W4289463412 workType "article" @default.