{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2025,11,7]],"date-time":"2025-11-07T13:22:55Z","timestamp":1762521775165},"reference-count":11,"publisher":"Association for Computing Machinery (ACM)","issue":"13","content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["Proc. VLDB Endow."],"published-print":{"date-parts":[[2014,8]]},"abstract":"<jats:p>Deep learning gains lots of attentions in recent years and is more and more important for mining values in big data. However, to make deep learning practical for a wide range of applications in Tencent Inc., three requirements must be considered: 1) Lots of computational power are required to train a practical model with tens of millions of parameters and billions of samples for products such as automatic speech recognition (ASR), and the number of parameters and training data is still growing. 2) The capability of training larger model is necessary for better model quality. 3) Easy to use frameworks are valuable to do many experiments to perform model selection, such as finding an appropriate optimization algorithm and tuning optimal hyper-parameters. To accelerate training, support large models, and make experiments easier, we built Mariana, the Tencent deep learning platform, which utilizes GPU and CPU cluster to train models parallelly with three frameworks: 1) a multi-GPU data parallelism framework for deep neural networks (DNNs). 2) a multi-GPU model parallelism and data parallelism framework for deep convolutional neural networks (CNNs). 3) a CPU cluster framework for large scale DNNs. Mariana also provides built-in algorithms and features to facilitate experiments. Mariana is in production usage for more than one year, achieves state-of-the-art acceleration performance, and plays a key role in training models and improving quality for automatic speech recognition and image recognition in Tencent WeChat, a mobile social platform, and for Ad click-through rate prediction (pCTR) in Tencent QQ, an instant messaging platform, and Tencent Qzone, a social networking service.<\/jats:p>","DOI":"10.14778\/2733004.2733082","type":"journal-article","created":{"date-parts":[[2015,5,12]],"date-time":"2015-05-12T15:37:52Z","timestamp":1431445072000},"page":"1772-1777","source":"Crossref","is-referenced-by-count":36,"title":["Mariana"],"prefix":"10.14778","volume":"7","author":[{"given":"Yongqiang","family":"Zou","sequence":"first","affiliation":[{"name":"Tencent Inc."}]},{"given":"Xing","family":"Jin","sequence":"additional","affiliation":[{"name":"Tencent Inc."}]},{"given":"Yi","family":"Li","sequence":"additional","affiliation":[{"name":"Tencent Inc."}]},{"given":"Zhimao","family":"Guo","sequence":"additional","affiliation":[{"name":"Tencent Inc."}]},{"given":"Eryu","family":"Wang","sequence":"additional","affiliation":[{"name":"Tencent Inc."}]},{"given":"Bin","family":"Xiao","sequence":"additional","affiliation":[{"name":"Tencent Inc."}]}],"member":"320","published-online":{"date-parts":[[2014,8]]},"reference":[{"key":"e_1_2_1_1_1","first-page":"6","volume":"313","author":"Hinton G. E.","year":"2006","unstructured":"Hinton , G. E. , and Salakhutdinov , R. R. Reducing the Dimensionality of Data with Neural Networks. Science , 313 , 6 ( July . 2006 ), 504--507. Hinton, G. E., and Salakhutdinov, R. R. Reducing the Dimensionality of Data with Neural Networks. Science, 313, 6 (July. 2006), 504--507.","journal-title":"Reducing the Dimensionality of Data with Neural Networks. Science"},{"key":"e_1_2_1_2_1","doi-asserted-by":"publisher","DOI":"10.1109\/MSP.2012.2205597"},{"key":"e_1_2_1_3_1","volume-title":"Proceedings of the Neural Information Processing Systems (NIPS'12)","author":"Krizhevsky A.","year":"2012","unstructured":"Krizhevsky , A. , Sutskever , I. , and Hinton , G.E . ImageNet Classification with Deep Convolutional Neural Networks . In Proceedings of the Neural Information Processing Systems (NIPS'12) (Lake Tahoe, Nevada, United States, December 3--6 , 2012 ). Curran Associates, Inc, 57 Morehouse Lane, Red Hook, NY, 2013, 1097--1106. Krizhevsky, A., Sutskever, I., and Hinton, G.E. ImageNet Classification with Deep Convolutional Neural Networks. In Proceedings of the Neural Information Processing Systems (NIPS'12) (Lake Tahoe, Nevada, United States, December 3--6, 2012). Curran Associates, Inc, 57 Morehouse Lane, Red Hook, NY, 2013, 1097--1106."},{"key":"e_1_2_1_4_1","volume-title":"A. Y. Large Scale Distributed Deep Networks. In Proceedings of the Neural Information Processing Systems (NIPS'12)","author":"Dean J.","year":"2012","unstructured":"Dean , J. , Corrado , G.S. , Monga , R. , Ng , A. Y. Large Scale Distributed Deep Networks. In Proceedings of the Neural Information Processing Systems (NIPS'12) (Lake Tahoe, Nevada, United States, December 3--6 , 2012 ). Curran Associates, Inc, 57 Morehouse Lane, Red Hook, NY, 2013, 1223--1232. Dean, J., Corrado, G.S., Monga, R., et al, Ng, A. Y. Large Scale Distributed Deep Networks. In Proceedings of the Neural Information Processing Systems (NIPS'12) (Lake Tahoe, Nevada, United States, December 3--6, 2012). Curran Associates, Inc, 57 Morehouse Lane, Red Hook, NY, 2013, 1223--1232."},{"key":"e_1_2_1_5_1","volume-title":"Proceedings of the 30th International Conference on Machine Learning (ICML'13)","volume":"2013","author":"Coates A.","year":"2013","unstructured":"Coates , A. , Huval , B. , Wang , T. , Wu , D. J. , Ng , A. Y. Deep learning with COTS HPC systems . In Proceedings of the 30th International Conference on Machine Learning (ICML'13) (Atlanta, Georgia, USA, June 16--21 , 2013 ). JMLR: W&CP volume 28(3), 2013 , 1337--1345. Coates, A., Huval, B., Wang, T., Wu, D. J., Ng, A. Y. Deep learning with COTS HPC systems. In Proceedings of the 30th International Conference on Machine Learning (ICML'13) (Atlanta, Georgia, USA, June 16--21, 2013). JMLR: W&CP volume 28(3), 2013, 1337--1345."},{"key":"e_1_2_1_6_1","series-title":"February 2014","volume-title":"Multi-GPU Training of ConvNets. arXiv:1312.5853v4 {cs.LG}","author":"Yadan O.","unstructured":"Yadan , O. , Adams , K. , Taigman , Y. , Ranzato , M. A. Multi-GPU Training of ConvNets. arXiv:1312.5853v4 {cs.LG} ( February 2014 ) Yadan, O., Adams, K., Taigman, Y., Ranzato, M. A. Multi-GPU Training of ConvNets. arXiv:1312.5853v4 {cs.LG} (February 2014)"},{"key":"e_1_2_1_7_1","first-page":"7","volume":"12","author":"Duchi J. C.","year":"2011","unstructured":"Duchi , J. C. , Hazan , E. , and Singer , Y. Adaptive subgradient methods for online learning and stochastic optimization. Journal of Machine Learning Research , 12 , 7 ( July . 2011 ), 2121--2159. Duchi, J. C., Hazan, E., and Singer, Y. Adaptive subgradient methods for online learning and stochastic optimization. Journal of Machine Learning Research, 12, 7 (July. 2011), 2121--2159.","journal-title":"Journal of Machine Learning Research"},{"key":"e_1_2_1_8_1","volume-title":"USA","author":"Krizhevsky A.","year":"2014","unstructured":"Krizhevsky , A. Parallelizing Convolutional Neural Networks. in tutorial of IEEE Conference on Computer Vision and Pattern Recognition (CVPR 2014). (Columbus, Ohio , USA , June 23-28, 2014 ). 2014. Krizhevsky, A. Parallelizing Convolutional Neural Networks. in tutorial of IEEE Conference on Computer Vision and Pattern Recognition (CVPR 2014). (Columbus, Ohio, USA, June 23-28, 2014). 2014."},{"key":"e_1_2_1_9_1","doi-asserted-by":"crossref","unstructured":"Jia Y. Q. Caffe: An Open Source Convolutional Architecture for Fast Feature Embedding. http:\/\/caffe.berkeleyvision.org (2013).  Jia Y. Q. Caffe: An Open Source Convolutional Architecture for Fast Feature Embedding. http:\/\/caffe.berkeleyvision.org (2013).","DOI":"10.1145\/2647868.2654889"},{"key":"e_1_2_1_10_1","volume-title":"The Kaldi Speech Recognition Toolkit. in Proceedings of IEEE 2011 Workshop on Automatic Speech Recognition and Understanding(ASRU 2011)","author":"Povey D.","year":"2011","unstructured":"Povey , D. , Ghoshal , A. Boulianne , G. , Vesely , K. Kaldi . The Kaldi Speech Recognition Toolkit. in Proceedings of IEEE 2011 Workshop on Automatic Speech Recognition and Understanding(ASRU 2011) (Hilton Waikoloa Village, Big Island, Hawaii, US , December 11-15, 2011 ). IEEE Signal Processing Society. IEEE Catalog No.: CFP11SRW-USB. Povey, D., Ghoshal, A. Boulianne, G., et al, Vesely, K. Kaldi. The Kaldi Speech Recognition Toolkit. in Proceedings of IEEE 2011 Workshop on Automatic Speech Recognition and Understanding(ASRU 2011) (Hilton Waikoloa Village, Big Island, Hawaii, US, December 11-15, 2011). IEEE Signal Processing Society. IEEE Catalog No.: CFP11SRW-USB."},{"key":"e_1_2_1_11_1","volume-title":"Proceedings of the Python for Scientific Computing Conference (SciPy 2010","author":"Bergstra J.","year":"2010","unstructured":"Bergstra , J. , Breuleux , O. , Bastien , F. , Bengio , Y. \" Theano : A CPU and GPU Math Expression Compiler \". in Proceedings of the Python for Scientific Computing Conference (SciPy 2010 ). (Austin, Texas, USA. June 30 - July 3, 2010 ). Bergstra, J., Breuleux, O., Bastien, F., et al, Bengio, Y. \"Theano: A CPU and GPU Math Expression Compiler\". in Proceedings of the Python for Scientific Computing Conference (SciPy 2010). (Austin, Texas, USA. June 30 - July 3, 2010)."}],"container-title":["Proceedings of the VLDB Endowment"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/dl.acm.org\/doi\/pdf\/10.14778\/2733004.2733082","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2022,12,28]],"date-time":"2022-12-28T09:45:01Z","timestamp":1672220701000},"score":1,"resource":{"primary":{"URL":"https:\/\/dl.acm.org\/doi\/10.14778\/2733004.2733082"}},"subtitle":["tencent deep learning platform and its applications"],"short-title":[],"issued":{"date-parts":[[2014,8]]},"references-count":11,"journal-issue":{"issue":"13","published-print":{"date-parts":[[2014,8]]}},"alternative-id":["10.14778\/2733004.2733082"],"URL":"https:\/\/doi.org\/10.14778\/2733004.2733082","relation":{},"ISSN":["2150-8097"],"issn-type":[{"value":"2150-8097","type":"print"}],"subject":[],"published":{"date-parts":[[2014,8]]}}}