A comparative study of open source deep learning frameworks

Deep Learning (DL) is one of the hottest trends in machine learning as DL approaches produced results superior to the state-of-the-art in problematic areas such as image processing and natural language processing (NLP). To foster the growth of the DL community, several open source frameworks appeared providing implementations of the most common DL algorithms. These frameworks vary in the algorithms they support and in the quality of their implementations. The purpose of this work is to provide a qualitative and quantitative comparison among three of the most popular and most comprehensive DL frameworks (namely Google's TensorFlow, University of Montreal's Theano, and Microsoft's CNTK). The ultimate goal of this work is to help end users make an informed decision about the best DL framework that suits their needs and resources. To ensure that our study is as comprehensive as possible, we conduct several experiments using multiple benchmark datasets and measure the performance of the frameworks' implementation of different DL algorithms. For most of our experiments, we find out that CNTK's implementations are superior to the other ones under consideration.

[1]  Christopher Potts,et al.  Learning Word Vectors for Sentiment Analysis , 2011, ACL.

[2]  方华.  google,我,萨娜 , 2006 .

[3]  Martín Abadi,et al.  TensorFlow: Large-Scale Machine Learning on Heterogeneous Distributed Systems , 2016, ArXiv.

[4]  A. Kalinovsky,et al.  Deep learning with theano, torch, caffe, tensorflow, and deeplearning4J: which one is the best in speed and accuracy? , 2016 .

[5]  Geoffrey E. Hinton,et al.  ImageNet classification with deep convolutional neural networks , 2012, Commun. ACM.

[6]  Peter Goldsborough,et al.  A Tour of TensorFlow , 2016, ArXiv.

[7]  Geoffrey Zweig,et al.  An introduction to computational networks and the computational network toolkit (invited talk) , 2014, INTERSPEECH.

[8]  Yu Zhang,et al.  The Computational Network Toolkit [Best of the Web] , 2015, IEEE Signal Process. Mag..

[9]  Guigang Zhang,et al.  Deep Learning , 2016, Int. J. Semantic Comput..

[10]  Alex Krizhevsky,et al.  Learning Multiple Layers of Features from Tiny Images , 2009 .

[11]  Mahmoud Al-Ayyoub,et al.  Deep learning for Arabic NLP: A survey , 2017, J. Comput. Sci..

[12]  Mohak Shah,et al.  Comparative Study of Deep Learning Software Frameworks , 2015, 1511.06435.

[13]  Razvan Pascanu,et al.  Theano: A CPU and GPU Math Compiler in Python , 2010, SciPy.

[14]  Yoshua Bengio,et al.  Gradient-based learning applied to document recognition , 1998, Proc. IEEE.

[15]  Qiang Wang,et al.  Benchmarking State-of-the-Art Deep Learning Software Tools , 2016, 2016 7th International Conference on Cloud Computing and Big Data (CCBD).

[16]  Jeff Johnson,et al.  Fast Convolutional Nets With fbfft: A GPU Performance Evaluation , 2014, ICLR.

[17]  John Salvatier,et al.  Theano: A Python framework for fast computation of mathematical expressions , 2016, ArXiv.

[18]  Yann LeCun,et al.  The mnist database of handwritten digits , 2005 .