--- title: "MobileNets: Efficient Convolutional Neural — Howard 2017" description: "Modern CNNs usually gain accuracy by becoming deeper and heavier." authors: ["Andrew G. Howard", "Menglong Zhu", "Bo Chen", "Dmitry Kalenichenko", "Weijun Wang", "Tobias Weyand", "Marco Andreetto", "Hartwig Adam"] year: 2017 venue: "arXiv:1704.04861" source: "https://arxiv.org/abs/1704.04861" doi: "10.48550/arxiv.1704.04861" arxiv: "1704.04861" licence: "arXiv — non-exclusive license to distribute" licence_url: "http://arxiv.org/licenses/nonexclusive-distrib/1.0/" summary_licence: "Elamur Cosmic License" canonical: "https://research.elamur.ai/free/en/papers/howard-2017-mobilenets-efficient-convolutional-neural-networks/" --- [Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/howard-2017-mobilenets-efficient-convolutional-neural-networks/) [home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2017](https://md.research.elamur.ai/free/en/papers/2017/page/3/) / MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications # MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications **Authors:** Andrew G. Howard, Menglong Zhu, Bo Chen, [Dmitry Kalenichenko](https://md.research.elamur.ai/free/en/authors/dmitry-kalenichenko/), Weijun Wang, Tobias Weyand, Marco Andreetto, Hartwig Adam · **Year:** 2017 · **Venue:** arXiv:1704.04861 · **Licence:** [arXiv — non-exclusive license to distribute](http://arxiv.org/licenses/nonexclusive-distrib/1.0/) **Links:** [Publisher's page](https://arxiv.org/abs/1704.04861) · [PDF](https://arxiv.org/pdf/1704.04861) · [DOI 10.48550/arxiv.1704.04861](https://doi.org/10.48550/arxiv.1704.04861) · [arXiv 1704.04861](https://arxiv.org/abs/1704.04861) · [OpenAlex](https://openalex.org/W2612445135) **Citations:** 9185 citations (openalex-snapshot · 2026-10-14) · [448](#cited-by) in our library, of the 140884 papers whose bibliographies we read ## Summary Modern CNNs usually gain accuracy by becoming deeper and heavier. That makes them hard to run on phones, robots or embedded hardware, where latency and memory are tight. This Google paper introduces MobileNets, a family of compact image models designed mainly for speed and also for small size. The core building block is the depthwise separable convolution. Each input channel is first filtered independently, and a 1 × 1 pointwise convolution then mixes the channels. Together these replace one expensive standard convolution. On top of this 28-layer base network, two global settings let developers fit a model to their budget: the width multiplier thins every layer, and the resolution multiplier shrinks the input image. Training used TensorFlow and RMSprop, with lighter regularization than large networks need. Beyond ImageNet classification, the authors test MobileNets on: - COCO object detection - Stanford Dogs fine-grained recognition - face attributes and face embeddings - PlaNet geolocalization Across these tasks, MobileNets typically lose a little accuracy while cutting computation and parameters many times over. This made capable vision models practical on-device. ## Key points - 3 × 3 depthwise separable convolutions need roughly 8–9× less computation than standard convolutions, with only a small accuracy loss. The 1 × 1 layers take about 95% of compute time and hold 75% of the parameters, and they map efficiently onto GEMM. - COCO detection (SSD 300): MobileNet reaches 19.3% mAP with 1.2B Mult-Adds and 6.8M parameters. Inception V2 reaches 22.0% with 3.8B Mult-Adds. - Face attributes via distillation: - 1.0 MobileNet-224 scores 88.7% mean AP at 568M Mult-Adds, versus 86.9% for the 1600M in-house baseline. - 0.25 MobileNet-128 still scores 86.4% at only 15M Mult-Adds. - PlaNet geolocalization: the MobileNet version has 13M parameters versus 52M for the Inception V3 original. Its accuracy drops only slightly, and it still clearly beats Im2GPS. - A FaceNet model distilled into 1.0 MobileNet-160 reaches 79.4% accuracy, versus 83% for FaceNet, using 286M instead of 1600M Mult-Adds. *Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.* Fields: [Other](https://md.research.elamur.ai/free/en/shelves/other/) ## Cited by Cited by 448 papers of our library cite it (of the 140884 whose bibliographies we read); 7 of them have a page here, 441 not yet. - [YMIR: A new Benchmark Dataset and Model for Arabic Yemeni Music Genre Classification Using Convolutional Neural Networks](https://md.research.elamur.ai/free/en/papers/al-2026-makhlafi-ymir-new-benchmark-dataset-model/) - [TVM: An Automated End-to-End Optimizing Compiler for Deep Learning](https://md.research.elamur.ai/free/en/papers/chen-2018-tvm-automated-end-end-optimizing/) - [Green AI](https://md.research.elamur.ai/free/en/papers/schwartz-2019-green-ai/) - [MobileBERT: a Compact Task-Agnostic BERT for Resource-Limited Devices](https://md.research.elamur.ai/free/en/papers/sun-2020-mobilebert-compact-task-agnostic-bert/) - [EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks](https://md.research.elamur.ai/free/en/papers/tan-2019-efficientnet-rethinking-model-scaling-convolutional/) - [Triton: An Intermediate Language and Compiler for Tiled Neural Network Computations](https://md.research.elamur.ai/free/en/papers/tillet-2019-triton-intermediate-language-compiler-tiled/) - [Learning Transferable Architectures for Scalable Image Recognition](https://md.research.elamur.ai/free/en/papers/zoph-2017-learning-transferable-architectures-scalable-image/) ## Cite The paper itself, not this page. BibTeX from the library's record of the work. ```bibtex @misc{howard2017mobilenets, title = {{MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications}}, author = {Howard, Andrew G. and Zhu, Menglong and Chen, Bo and Kalenichenko, Dmitry and Wang, Weijun and Weyand, Tobias and Andreetto, Marco and Adam, Hartwig}, year = {2017}, doi = {10.48550/arxiv.1704.04861}, eprint = {1704.04861}, archivePrefix = {arXiv}, url = {https://doi.org/10.48550/arxiv.1704.04861} } ``` - [← previous mixup: Beyond Empirical Risk Minimization](https://md.research.elamur.ai/free/en/papers/zhang-2017-mixup-empirical-risk-minimization/) - [next → Morphological Analysis for the Maltese Language: The challenges of a hybrid system](https://md.research.elamur.ai/free/en/papers/aclbulk-2017-borg-morphological-analysis-maltese-language-challenges/) - [↑ back to papers of 2017](https://md.research.elamur.ai/free/en/papers/2017/page/3/) ## References 37 references, 6 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet. 1. M. Abadi, A. Agarwal, P. Barham, E. Brevdo, Z. Chen, C. Citro, G. S. Corrado, A. Davis, J. Dean, M. Devin, et al. Tensorflow: Large-scale machine learning on heterogeneous systems, 2015. Software available from tensorflow. org, 1, 2015. 4 [find](https://scholar.google.com/scholar?q=M%2E+Abadi%2C+A%2E+Agarwal%2C+P%2E+Barham%2C+E%2E+Brevdo%2C+Z%2E+Chen%2C+C%2E+Citro%2C+G%2E+S%2E+Corrado%2C+A%2E+Davis%2C+J%2E+Dean%2C+M%2E+Devin%2C+et+al%2E+Tensorflow%3A+Large%2Dscale+machine+learning+on+heterogeneous+systems%2C+2015%2E+Software+avai) 2. W. Chen, J. T. Wilson, S. Tyree, K. Q. Weinberger, and Y. Chen. Compressing neural networks with the hashing trick. CoRR, abs/1504.04788, 2015. 2 — [arXiv:1504.04788](https://arxiv.org/abs/1504.04788) [find](https://scholar.google.com/scholar?q=W%2E+Chen%2C+J%2E+T%2E+Wilson%2C+S%2E+Tyree%2C+K%2E+Q%2E+Weinberger%2C+and+Y%2E+Chen%2E+Compressing+neural+networks+with+the+hashing+trick%2E+CoRR%2C+abs%2F1504%2E04788%2C+2015%2E+2+%E2%80%94+%5BarXiv%3A1504%2E04788%5D%28https%3A%2F%2Farxiv%2Eorg%2Fabs%2F1504%2E04788%29) 3. F. Chollet. Xception: Deep learning with depthwise separable convolutions. arXiv preprint arXiv:1610.02357v2, 2016. 1 — [arXiv:1610.02357](https://arxiv.org/abs/1610.02357) [find](https://scholar.google.com/scholar?q=F%2E+Chollet%2E+Xception%3A+Deep+learning+with+depthwise+separable+convolutions%2E+arXiv+preprint+arXiv%3A1610%2E02357v2%2C+2016%2E+1+%E2%80%94+%5BarXiv%3A1610%2E02357%5D%28https%3A%2F%2Farxiv%2Eorg%2Fabs%2F1610%2E02357%29) 4. M. Courbariaux, J.-P. David, and Y. Bengio. Training deep neural networks with low precision multiplications. arXiv preprint arXiv:1412.7024, 2014. 2 — [arXiv:1412.7024](https://arxiv.org/abs/1412.7024) [find](https://scholar.google.com/scholar?q=M%2E+Courbariaux%2C+J%2E%2DP%2E+David%2C+and+Y%2E+Bengio%2E+Training+deep+neural+networks+with+low+precision+multiplications%2E+arXiv+preprint+arXiv%3A1412%2E7024%2C+2014%2E+2+%E2%80%94+%5BarXiv%3A1412%2E7024%5D%28https%3A%2F%2Farxiv%2Eorg%2Fabs%2F1412%2E702) 5. S. Han, H. Mao, and W. J. Dally. Deep compression: Compressing deep neural network with pruning, trained quantization and huffman coding. CoRR, abs/1510.00149, 2, 2015. 2 [find](https://scholar.google.com/scholar?q=S%2E+Han%2C+H%2E+Mao%2C+and+W%2E+J%2E+Dally%2E+Deep+compression%3A+Compressing+deep+neural+network+with+pruning%2C+trained+quantization+and+huffman+coding%2E+CoRR%2C+abs%2F1510%2E00149%2C+2%2C+2015%2E+2) 6. J. Hays and A. Efros. IM2GPS: estimating geographic information from a single image. In Proceedings of the IEEE International Conference on Computer Vision and Pattern Recognition, 2008. 7 [find](https://scholar.google.com/scholar?q=J%2E+Hays+and+A%2E+Efros%2E+IM2GPS%3A+estimating+geographic+information+from+a+single+image%2E+In+Proceedings+of+the+IEEE+International+Conference+on+Computer+Vision+and+Pattern+Recognition%2C+2008%2E+7) 7. J. Hays and A. Efros. Large-Scale Image Geolocalization. In J. Choi and G. Friedland, editors, Multimodal Location Estimation of Videos and Images. Springer, 2014. 6, 7 [find](https://scholar.google.com/scholar?q=J%2E+Hays+and+A%2E+Efros%2E+Large%2DScale+Image+Geolocalization%2E+In+J%2E+Choi+and+G%2E+Friedland%2C+editors%2C+Multimodal+Location+Estimation+of+Videos+and+Images%2E+Springer%2C+2014%2E+6%2C+7) 8. K. He, X. Zhang, S. Ren, and J. Sun. Deep residual learning for image recognition. arXiv preprint arXiv:1512.03385, 2015. 1 [find](https://scholar.google.com/scholar?q=K%2E+He%2C+X%2E+Zhang%2C+S%2E+Ren%2C+and+J%2E+Sun%2E+Deep+residual+learning+for+image+recognition%2E+arXiv+preprint+arXiv%3A1512%2E03385%2C+2015%2E+1) 9. G. Hinton, O. Vinyals, and J. Dean. Distilling the knowledge in a neural network. arXiv preprint arXiv:1503.02531, 2015. 2, 7 [find](https://scholar.google.com/scholar?q=G%2E+Hinton%2C+O%2E+Vinyals%2C+and+J%2E+Dean%2E+Distilling+the+knowledge+in+a+neural+network%2E+arXiv+preprint+arXiv%3A1503%2E02531%2C+2015%2E+2%2C+7) [our card](https://md.research.elamur.ai/free/en/papers/hinton-2015-distilling-knowledge-neural-network/) 10. J. Huang, V. Rathod, C. Sun, M. Zhu, A. Korattikara, A. Fathi, I. Fischer, Z. Wojna, Y. Song, S. Guadarrama, et al. Speed/accuracy trade-offs for modern convolutional object detectors. arXiv preprint arXiv:1611.10012, 2016. 7 — [arXiv:1611.10012](https://arxiv.org/abs/1611.10012) [find](https://scholar.google.com/scholar?q=J%2E+Huang%2C+V%2E+Rathod%2C+C%2E+Sun%2C+M%2E+Zhu%2C+A%2E+Korattikara%2C+A%2E+Fathi%2C+I%2E+Fischer%2C+Z%2E+Wojna%2C+Y%2E+Song%2C+S%2E+Guadarrama%2C+et+al%2E+Speed%2Faccuracy+trade%2Doffs+for+modern+convolutional+object+detectors%2E+arXiv+preprint+) 11. I. Hubara, M. Courbariaux, D. Soudry, R. El-Yaniv, and Y. Bengio. Quantized neural networks: Training neural networks with low precision weights and activations. arXiv preprint arXiv:1609.07061, 2016. 2 [find](https://scholar.google.com/scholar?q=I%2E+Hubara%2C+M%2E+Courbariaux%2C+D%2E+Soudry%2C+R%2E+El%2DYaniv%2C+and+Y%2E+Bengio%2E+Quantized+neural+networks%3A+Training+neural+networks+with+low+precision+weights+and+activations%2E+arXiv+preprint+arXiv%3A1609%2E07061%2C+2016%2E) 12. F. N. Iandola, M. W. Moskewicz, K. Ashraf, S. Han, W. J. Dally, and K. Keutzer. Squeezenet: Alexnet-level accuracy with 50x fewer parameters and¡ 1mb model size. arXiv preprint arXiv:1602.07360, 2016. 1, 6 — [arXiv:1602.07360](https://arxiv.org/abs/1602.07360) [find](https://scholar.google.com/scholar?q=F%2E+N%2E+Iandola%2C+M%2E+W%2E+Moskewicz%2C+K%2E+Ashraf%2C+S%2E+Han%2C+W%2E+J%2E+Dally%2C+and+K%2E+Keutzer%2E+Squeezenet%3A+Alexnet%2Dlevel+accuracy+with+50x+fewer+parameters+and%C2%A1+1mb+model+size%2E+arXiv+preprint+arXiv%3A1602%2E07360%2C+2016%2E) 13. S. Ioffe and C. Szegedy. Batch normalization: Accelerating deep network training by reducing internal covariate shift. arXiv preprint arXiv:1502.03167, 2015. 1, 3, 7 [find](https://scholar.google.com/scholar?q=S%2E+Ioffe+and+C%2E+Szegedy%2E+Batch+normalization%3A+Accelerating+deep+network+training+by+reducing+internal+covariate+shift%2E+arXiv+preprint+arXiv%3A1502%2E03167%2C+2015%2E+1%2C+3%2C+7) 14. M. Jaderberg, A. Vedaldi, and A. Zisserman. Speeding up convolutional neural networks with low rank expansions. arXiv preprint arXiv:1405.3866, 2014. 2 — [arXiv:1405.3866](https://arxiv.org/abs/1405.3866) [find](https://scholar.google.com/scholar?q=M%2E+Jaderberg%2C+A%2E+Vedaldi%2C+and+A%2E+Zisserman%2E+Speeding+up+convolutional+neural+networks+with+low+rank+expansions%2E+arXiv+preprint+arXiv%3A1405%2E3866%2C+2014%2E+2+%E2%80%94+%5BarXiv%3A1405%2E3866%5D%28https%3A%2F%2Farxiv%2Eorg%2Fabs%2F1405%2E3) 15. Y. Jia, E. Shelhamer, J. Donahue, S. Karayev, J. Long, R. Girshick, S. Guadarrama, and T. Darrell. Caffe: Convolutional architecture for fast feature embedding. arXiv preprint arXiv:1408.5093, 2014. 4 [find](https://scholar.google.com/scholar?q=Y%2E+Jia%2C+E%2E+Shelhamer%2C+J%2E+Donahue%2C+S%2E+Karayev%2C+J%2E+Long%2C+R%2E+Girshick%2C+S%2E+Guadarrama%2C+and+T%2E+Darrell%2E+Caffe%3A+Convolutional+architecture+for+fast+feature+embedding%2E+arXiv+preprint+arXiv%3A1408%2E5093%2C+2014%2E+4) [our card](https://md.research.elamur.ai/free/en/papers/jia-2014-caffe-convolutional-architecture-fast-feature/) 16. J. Jin, A. Dundar, and E. Culurciello. Flattened convolutional neural networks for feedforward acceleration. arXiv preprint arXiv:1412.5474, 2014. 1, 3 — [arXiv:1412.5474](https://arxiv.org/abs/1412.5474) [find](https://scholar.google.com/scholar?q=J%2E+Jin%2C+A%2E+Dundar%2C+and+E%2E+Culurciello%2E+Flattened+convolutional+neural+networks+for+feedforward+acceleration%2E+arXiv+preprint+arXiv%3A1412%2E5474%2C+2014%2E+1%2C+3+%E2%80%94+%5BarXiv%3A1412%2E5474%5D%28https%3A%2F%2Farxiv%2Eorg%2Fabs%2F1412%2E5) 17. A. Khosla, N. Jayadevaprakash, B. Yao, and L. Fei-Fei. Novel dataset for fine-grained image categorization. In First Workshop on Fine-Grained Visual Categorization, IEEE Conference on Computer Vision and Pattern Recognition, Colorado Springs, CO, June 2011. 6 [find](https://scholar.google.com/scholar?q=A%2E+Khosla%2C+N%2E+Jayadevaprakash%2C+B%2E+Yao%2C+and+L%2E+Fei%2DFei%2E+Novel+dataset+for+fine%2Dgrained+image+categorization%2E+In+First+Workshop+on+Fine%2DGrained+Visual+Categorization%2C+IEEE+Conference+on+Computer+Vision+) 18. J. Krause, B. Sapp, A. Howard, H. Zhou, A. Toshev, T. Duerig, J. Philbin, and L. Fei-Fei. The unreasonable effectiveness of noisy data for fine-grained recognition. arXiv preprint arXiv:1511.06789, 2015. 6 — [arXiv:1511.06789](https://arxiv.org/abs/1511.06789) [find](https://scholar.google.com/scholar?q=J%2E+Krause%2C+B%2E+Sapp%2C+A%2E+Howard%2C+H%2E+Zhou%2C+A%2E+Toshev%2C+T%2E+Duerig%2C+J%2E+Philbin%2C+and+L%2E+Fei%2DFei%2E+The+unreasonable+effectiveness+of+noisy+data+for+fine%2Dgrained+recognition%2E+arXiv+preprint+arXiv%3A1511%2E06789%2C+20) 19. A. Krizhevsky, I. Sutskever, and G. E. Hinton. Imagenet classification with deep convolutional neural networks. In Advances in neural information processing systems, pages 1097–1105, 2012. 1, 6 [find](https://scholar.google.com/scholar?q=A%2E+Krizhevsky%2C+I%2E+Sutskever%2C+and+G%2E+E%2E+Hinton%2E+Imagenet+classification+with+deep+convolutional+neural+networks%2E+In+Advances+in+neural+information+processing+systems%2C+pages+1097%E2%80%931105%2C+2012%2E+1%2C+6) 20. V. Lebedev, Y. Ganin, M. Rakhuba, I. Oseledets, and V. Lempitsky. Speeding-up convolutional neural networks using fine-tuned cp-decomposition. arXiv preprint arXiv:1412.6553, 2014. 2 — [arXiv:1412.6553](https://arxiv.org/abs/1412.6553) [find](https://scholar.google.com/scholar?q=V%2E+Lebedev%2C+Y%2E+Ganin%2C+M%2E+Rakhuba%2C+I%2E+Oseledets%2C+and+V%2E+Lempitsky%2E+Speeding%2Dup+convolutional+neural+networks+using+fine%2Dtuned+cp%2Ddecomposition%2E+arXiv+preprint+arXiv%3A1412%2E6553%2C+2014%2E+2+%E2%80%94+%5BarXiv%3A1412%2E655) 21. W. Liu, D. Anguelov, D. Erhan, C. Szegedy, and S. Reed. Ssd: Single shot multibox detector. arXiv preprint arXiv:1512.02325, 2015. 7 — [arXiv:1512.02325](https://arxiv.org/abs/1512.02325) [find](https://scholar.google.com/scholar?q=W%2E+Liu%2C+D%2E+Anguelov%2C+D%2E+Erhan%2C+C%2E+Szegedy%2C+and+S%2E+Reed%2E+Ssd%3A+Single+shot+multibox+detector%2E+arXiv+preprint+arXiv%3A1512%2E02325%2C+2015%2E+7+%E2%80%94+%5BarXiv%3A1512%2E02325%5D%28https%3A%2F%2Farxiv%2Eorg%2Fabs%2F1512%2E02325%29) 22. M. Rastegari, V. Ordonez, J. Redmon, and A. Farhadi. Xnor-net: Imagenet classification using binary convolutional neural networks. arXiv preprint arXiv:1603.05279, 2016. 1, 2 [find](https://scholar.google.com/scholar?q=M%2E+Rastegari%2C+V%2E+Ordonez%2C+J%2E+Redmon%2C+and+A%2E+Farhadi%2E+Xnor%2Dnet%3A+Imagenet+classification+using+binary+convolutional+neural+networks%2E+arXiv+preprint+arXiv%3A1603%2E05279%2C+2016%2E+1%2C+2) 23. S. Ren, K. He, R. Girshick, and J. Sun. Faster r-cnn: Towards real-time object detection with region proposal networks. In Advances in neural information processing systems, pages 91–99, 2015. 7 [find](https://scholar.google.com/scholar?q=S%2E+Ren%2C+K%2E+He%2C+R%2E+Girshick%2C+and+J%2E+Sun%2E+Faster+r%2Dcnn%3A+Towards+real%2Dtime+object+detection+with+region+proposal+networks%2E+In+Advances+in+neural+information+processing+systems%2C+pages+91%E2%80%9399%2C+2015%2E+7) [our card](https://md.research.elamur.ai/free/en/papers/ren-2015-faster-r-cnn-towards-real/) 24. O. Russakovsky, J. Deng, H. Su, J. Krause, S. Satheesh, S. Ma, Z. Huang, A. Karpathy, A. Khosla, M. Bernstein, et al. Imagenet large scale visual recognition challenge. International Journal of Computer Vision, 115(3):211–252, 2015. 1 [find](https://scholar.google.com/scholar?q=O%2E+Russakovsky%2C+J%2E+Deng%2C+H%2E+Su%2C+J%2E+Krause%2C+S%2E+Satheesh%2C+S%2E+Ma%2C+Z%2E+Huang%2C+A%2E+Karpathy%2C+A%2E+Khosla%2C+M%2E+Bernstein%2C+et+al%2E+Imagenet+large+scale+visual+recognition+challenge%2E+International+Journal+of+Comput) [our card](https://md.research.elamur.ai/free/en/papers/russakovsky-2014-imagenet-large-scale-visual-recognition/) 25. F. Schroff, D. Kalenichenko, and J. Philbin. Facenet: A unified embedding for face recognition and clustering. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pages 815–823, 2015. 8 [find](https://scholar.google.com/scholar?q=F%2E+Schroff%2C+D%2E+Kalenichenko%2C+and+J%2E+Philbin%2E+Facenet%3A+A+unified+embedding+for+face+recognition+and+clustering%2E+In+Proceedings+of+the+IEEE+Conference+on+Computer+Vision+and+Pattern+Recognition%2C+pages+8) [our card](https://md.research.elamur.ai/free/en/papers/schroff-2015-facenet-unified-embedding-face-recognition/) 26. L. Sifre. Rigid-motion scattering for image classification. PhD thesis, Ph. D. thesis, 2014. 1, 3 [find](https://scholar.google.com/scholar?q=L%2E+Sifre%2E+Rigid%2Dmotion+scattering+for+image+classification%2E+PhD+thesis%2C+Ph%2E+D%2E+thesis%2C+2014%2E+1%2C+3) 27. K. Simonyan and A. Zisserman. Very deep convolutional networks for large-scale image recognition. arXiv preprint arXiv:1409.1556, 2014. 1, 6 [find](https://scholar.google.com/scholar?q=K%2E+Simonyan+and+A%2E+Zisserman%2E+Very+deep+convolutional+networks+for+large%2Dscale+image+recognition%2E+arXiv+preprint+arXiv%3A1409%2E1556%2C+2014%2E+1%2C+6) [our card](https://md.research.elamur.ai/free/en/papers/simonyan-2014-very-deep-convolutional-networks-large/) 28. V. Sindhwani, T. Sainath, and S. Kumar. Structured transforms for small-footprint deep learning. In Advances in Neural Information Processing Systems, pages 3088–3096, 2015. 1 [find](https://scholar.google.com/scholar?q=V%2E+Sindhwani%2C+T%2E+Sainath%2C+and+S%2E+Kumar%2E+Structured+transforms+for+small%2Dfootprint+deep+learning%2E+In+Advances+in+Neural+Information+Processing+Systems%2C+pages+3088%E2%80%933096%2C+2015%2E+1) 29. C. Szegedy, S. Ioffe, and V. Vanhoucke. Inception-v4, inception-resnet and the impact of residual connections on learning. arXiv preprint arXiv:1602.07261, 2016. 1 — [arXiv:1602.07261](https://arxiv.org/abs/1602.07261) [find](https://scholar.google.com/scholar?q=C%2E+Szegedy%2C+S%2E+Ioffe%2C+and+V%2E+Vanhoucke%2E+Inception%2Dv4%2C+inception%2Dresnet+and+the+impact+of+residual+connections+on+learning%2E+arXiv+preprint+arXiv%3A1602%2E07261%2C+2016%2E+1+%E2%80%94+%5BarXiv%3A1602%2E07261%5D%28https%3A%2F%2Farxiv%2Eo) 30. C. Szegedy, W. Liu, Y. Jia, P. Sermanet, S. Reed, D. Anguelov, D. Erhan, V. Vanhoucke, and A. Rabinovich. Going deeper with convolutions. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pages 1–9, 2015. 6 [find](https://scholar.google.com/scholar?q=C%2E+Szegedy%2C+W%2E+Liu%2C+Y%2E+Jia%2C+P%2E+Sermanet%2C+S%2E+Reed%2C+D%2E+Anguelov%2C+D%2E+Erhan%2C+V%2E+Vanhoucke%2C+and+A%2E+Rabinovich%2E+Going+deeper+with+convolutions%2E+In+Proceedings+of+the+IEEE+Conference+on+Computer+Vision+and+P) 31. C. Szegedy, V. Vanhoucke, S. Ioffe, J. Shlens, and Z. Wojna. Rethinking the inception architecture for computer vision. arXiv preprint arXiv:1512.00567, 2015. 1, 3, 4, 7 [find](https://scholar.google.com/scholar?q=C%2E+Szegedy%2C+V%2E+Vanhoucke%2C+S%2E+Ioffe%2C+J%2E+Shlens%2C+and+Z%2E+Wojna%2E+Rethinking+the+inception+architecture+for+computer+vision%2E+arXiv+preprint+arXiv%3A1512%2E00567%2C+2015%2E+1%2C+3%2C+4%2C+7) 32. B. Thomee, D. A. Shamma, G. Friedland, B. Elizalde, K. Ni, D. Poland, D. Borth, and L.-J. Li. Yfcc100m: The new data in multimedia research. Communications of the ACM, 59(2):64–73, 2016. 7 [find](https://scholar.google.com/scholar?q=B%2E+Thomee%2C+D%2E+A%2E+Shamma%2C+G%2E+Friedland%2C+B%2E+Elizalde%2C+K%2E+Ni%2C+D%2E+Poland%2C+D%2E+Borth%2C+and+L%2E%2DJ%2E+Li%2E+Yfcc100m%3A+The+new+data+in+multimedia+research%2E+Communications+of+the+ACM%2C+59%282%29%3A64%E2%80%9373%2C+2016%2E+7) 33. T. Tieleman and G. Hinton. Lecture 6.5-rmsprop: Divide the gradient by a running average of its recent magnitude. COURSERA: Neural Networks for Machine Learning, 4(2), 2012. 4 [find](https://scholar.google.com/scholar?q=T%2E+Tieleman+and+G%2E+Hinton%2E+Lecture+6%2E5%2Drmsprop%3A+Divide+the+gradient+by+a+running+average+of+its+recent+magnitude%2E+COURSERA%3A+Neural+Networks+for+Machine+Learning%2C+4%282%29%2C+2012%2E+4) 34. M. Wang, B. Liu, and H. Foroosh. Factorized convolutional neural networks. arXiv preprint arXiv:1608.04337, 2016. 1 — [arXiv:1608.04337](https://arxiv.org/abs/1608.04337) [find](https://scholar.google.com/scholar?q=M%2E+Wang%2C+B%2E+Liu%2C+and+H%2E+Foroosh%2E+Factorized+convolutional+neural+networks%2E+arXiv+preprint+arXiv%3A1608%2E04337%2C+2016%2E+1+%E2%80%94+%5BarXiv%3A1608%2E04337%5D%28https%3A%2F%2Farxiv%2Eorg%2Fabs%2F1608%2E04337%29) 35. T. Weyand, I. Kostrikov, and J. Philbin. PlaNet - Photo Geolocation with Convolutional Neural Networks. In European Conference on Computer Vision (ECCV), 2016. 6, 7 [find](https://scholar.google.com/scholar?q=T%2E+Weyand%2C+I%2E+Kostrikov%2C+and+J%2E+Philbin%2E+PlaNet+%2D+Photo+Geolocation+with+Convolutional+Neural+Networks%2E+In+European+Conference+on+Computer+Vision+%28ECCV%29%2C+2016%2E+6%2C+7) 36. J. Wu, C. Leng, Y. Wang, Q. Hu, and J. Cheng. Quantized convolutional neural networks for mobile devices. arXiv preprint arXiv:1512.06473, 2015. 1 — [arXiv:1512.06473](https://arxiv.org/abs/1512.06473) [find](https://scholar.google.com/scholar?q=J%2E+Wu%2C+C%2E+Leng%2C+Y%2E+Wang%2C+Q%2E+Hu%2C+and+J%2E+Cheng%2E+Quantized+convolutional+neural+networks+for+mobile+devices%2E+arXiv+preprint+arXiv%3A1512%2E06473%2C+2015%2E+1+%E2%80%94+%5BarXiv%3A1512%2E06473%5D%28https%3A%2F%2Farxiv%2Eorg%2Fabs%2F1512%2E06473) 37. Z. Yang, M. Moczulski, M. Denil, N. de Freitas, A. Smola, L. Song, and Z. Wang. Deep fried convnets. In Proceedings of the IEEE International Conference on Computer Vision, pages 1476–1483, 2015. 1 [find](https://scholar.google.com/scholar?q=Z%2E+Yang%2C+M%2E+Moczulski%2C+M%2E+Denil%2C+N%2E+de+Freitas%2C+A%2E+Smola%2C+L%2E+Song%2C+and+Z%2E+Wang%2E+Deep+fried+convnets%2E+In+Proceedings+of+the+IEEE+International+Conference+on+Computer+Vision%2C+pages+1476%E2%80%931483%2C+2015%2E+1)