Matches in SemOpenAlex for { <https://semopenalex.org/work/W3044091840> ?p ?o ?g. }
Showing items 1 to 78 of
78
with 100 items per page.
- W3044091840 abstract "Designing convolutional neural networks (CNN) for mobile devices is challenging because mobile models need to be small and fast, yet still accurate. Although significant efforts have been dedicated to design and improve mobile CNNs on all dimensions, it is very difficult to manually balance these trade-offs when there are so many architectural possibilities to consider. In this paper, we propose an automated mobile neural architecture search (MNAS) approach, which explicitly incorporate model latency into the main objective so that the search can identify a model that achieves a good trade-off between accuracy and latency. Unlike previous work, where latency is considered via another, often inaccurate proxy (e.g., FLOPS), our approach directly measures real-world inference latency by executing the model on mobile phones. To further strike the right balance between flexibility and search space size, we propose a novel factorized hierarchical search space that encourages layer diversity throughout the network. Experimental results show that our approach consistently outperforms state-of-the-art mobile CNN models across multiple vision tasks. On the ImageNet classification task, our MnasNet achieves 75.2% top-1 accuracy with 78ms latency on a Pixel phone, which is 1.8x faster than MobileNetV2 [29] with 0.5% higher accuracy and 2.3x faster than NASNet [36] with 1.2% higher accuracy. Our MnasNet also achieves better mAP quality than MobileNets for COCO object detection. Code is at https://github.com/tensorflow/tpu/tree/master/models/official/mnasnet" @default.
- W3044091840 created "2020-07-29" @default.
- W3044091840 creator A5038861110 @default.
- W3044091840 creator A5048395030 @default.
- W3044091840 creator A5059366488 @default.
- W3044091840 creator A5060530625 @default.
- W3044091840 creator A5070063070 @default.
- W3044091840 creator A5075033315 @default.
- W3044091840 creator A5088927071 @default.
- W3044091840 date "2018-07-30" @default.
- W3044091840 modified "2023-09-23" @default.
- W3044091840 title "MnasNet: Platform-Aware Neural Architecture Search for Mobile" @default.
- W3044091840 doi "https://doi.org/10.48550/arxiv.1807.11626" @default.
- W3044091840 hasPublicationYear "2018" @default.
- W3044091840 type Work @default.
- W3044091840 sameAs 3044091840 @default.
- W3044091840 citedByCount "2" @default.
- W3044091840 countsByYear W30440918402023 @default.
- W3044091840 crossrefType "posted-content" @default.
- W3044091840 hasAuthorship W3044091840A5038861110 @default.
- W3044091840 hasAuthorship W3044091840A5048395030 @default.
- W3044091840 hasAuthorship W3044091840A5059366488 @default.
- W3044091840 hasAuthorship W3044091840A5060530625 @default.
- W3044091840 hasAuthorship W3044091840A5070063070 @default.
- W3044091840 hasAuthorship W3044091840A5075033315 @default.
- W3044091840 hasAuthorship W3044091840A5088927071 @default.
- W3044091840 hasBestOaLocation W30440918401 @default.
- W3044091840 hasConcept C111919701 @default.
- W3044091840 hasConcept C113775141 @default.
- W3044091840 hasConcept C119857082 @default.
- W3044091840 hasConcept C123657996 @default.
- W3044091840 hasConcept C142362112 @default.
- W3044091840 hasConcept C153349607 @default.
- W3044091840 hasConcept C154945302 @default.
- W3044091840 hasConcept C173608175 @default.
- W3044091840 hasConcept C186967261 @default.
- W3044091840 hasConcept C2776214188 @default.
- W3044091840 hasConcept C2984842247 @default.
- W3044091840 hasConcept C3826847 @default.
- W3044091840 hasConcept C41008148 @default.
- W3044091840 hasConcept C50644808 @default.
- W3044091840 hasConcept C76155785 @default.
- W3044091840 hasConcept C81363708 @default.
- W3044091840 hasConcept C82876162 @default.
- W3044091840 hasConceptScore W3044091840C111919701 @default.
- W3044091840 hasConceptScore W3044091840C113775141 @default.
- W3044091840 hasConceptScore W3044091840C119857082 @default.
- W3044091840 hasConceptScore W3044091840C123657996 @default.
- W3044091840 hasConceptScore W3044091840C142362112 @default.
- W3044091840 hasConceptScore W3044091840C153349607 @default.
- W3044091840 hasConceptScore W3044091840C154945302 @default.
- W3044091840 hasConceptScore W3044091840C173608175 @default.
- W3044091840 hasConceptScore W3044091840C186967261 @default.
- W3044091840 hasConceptScore W3044091840C2776214188 @default.
- W3044091840 hasConceptScore W3044091840C2984842247 @default.
- W3044091840 hasConceptScore W3044091840C3826847 @default.
- W3044091840 hasConceptScore W3044091840C41008148 @default.
- W3044091840 hasConceptScore W3044091840C50644808 @default.
- W3044091840 hasConceptScore W3044091840C76155785 @default.
- W3044091840 hasConceptScore W3044091840C81363708 @default.
- W3044091840 hasConceptScore W3044091840C82876162 @default.
- W3044091840 hasLocation W30440918401 @default.
- W3044091840 hasOpenAccess W3044091840 @default.
- W3044091840 hasPrimaryLocation W30440918401 @default.
- W3044091840 hasRelatedWork W2784537340 @default.
- W3044091840 hasRelatedWork W3015045227 @default.
- W3044091840 hasRelatedWork W3026990304 @default.
- W3044091840 hasRelatedWork W3195610113 @default.
- W3044091840 hasRelatedWork W3203251378 @default.
- W3044091840 hasRelatedWork W3205684019 @default.
- W3044091840 hasRelatedWork W4287822910 @default.
- W3044091840 hasRelatedWork W4297798330 @default.
- W3044091840 hasRelatedWork W4308075506 @default.
- W3044091840 hasRelatedWork W2943923678 @default.
- W3044091840 isParatext "false" @default.
- W3044091840 isRetracted "false" @default.
- W3044091840 magId "3044091840" @default.
- W3044091840 workType "article" @default.