Comparison of deep learning software

The following table compares some of the most popular software frameworks, libraries and computer programs for deep learning.

Deep learning software by name

This list is incomplete; you can help by expanding it.

Software	Creator	Software license[lower-alpha 1]	Open source	Platform	Written in	Interface	OpenMP support	OpenCL support	CUDA support	Has pretrained models	Recurrent Nets	Convolutional Nets	RBM/DBNs
Caffe	Berkeley Vision and Learning Center, community contributors	BSD 2-Clause License	Yes	Ubuntu, OS X, AWS,[1] unofficial Android port,[2] Windows support by Microsoft Research,[3] unofficial Windows port[4]	C++, Python[5]	C++, command line, Python, MATLAB[6]	No	Branch,[7] pull request,[8] third party implementation[9]	Yes	Yes[10]	Yes	Yes	No[11]
CNTK	Microsoft	Free[12]	Yes	Windows, Linux[13]	C++	Command line;[14] C++, Python and .NET interfaces coming[15]	Yes[16]	No	Yes	No	Yes[17]	Yes[17]	?[18]
Deeplearning4j	Various; originally Adam Gibson	Apache 2.0	Yes	Linux, OSX, Windows, Android, CyanogenMod (Cross-platform)	Java, Scala, C	Java, Scala, Clojure	?	No[19]	Yes[20]	Yes[21]	Yes	Yes	Yes
Neural Designer	Artelnics	Proprietary	No	Windows, OS X, Linux	C++	Graphical user interface	Yes	No	No	?	No	No	No
OpenNN	Artelnics	GNU LGPL	Yes	Cross platform	C++	C++	Yes	No	No	?	No	No	No
SINGA[22]	Apache Incubator	Apache 2.0	Yes	Linux	C++, Python	Python, C++	No	No	Yes	No	Yes	Yes	Yes
SystemML[23]	IBM Research, Databricks, Netflix[24]	Apache 2.0	Yes	Linux, Mac OS, Windows	Java, R	?	?	?	?	?	?	?	?
TensorFlow	Google Brain team	Apache 2.0	Yes	Linux, Mac OS X (no support for Windows yet[25])	C++, Python	Python, C/C++	No	No[26]	Yes	No	Yes	Yes	Yes
Theano	Université de Montréal	BSD license	Yes	Cross-platform	Python	Python	Yes	Under development[27]	Yes	Through Lasagne's model zoo[28]	Yes	Yes	Yes
Torch	Ronan Collobert, Koray Kavukcuoglu, Clement Farabet	BSD License	Yes	Linux, Android,[29] Mac OS X, iOS	C, Lua	Torch, C, utility library for C++/OpenCL[30]	Yes	Third party implementations[31]	Third party implementations[32]	Yes[33]	Yes	Yes	Yes

1. Licenses here are a summary, and are not taken to be complete statements of the licenses. Some libraries may use other libraries internally under different licenses

Deep learning software not yet covered

• adnn – Javascript neural networks
• Chainer – Neural network framework for Python
• cuDNN – Highly optimized deep learning computation primitives implemented in CUDA
• DeepX – Software accelerator for deep learning execution aimed towards mobile devices
• Faster RNNLM (HS/NCE) toolkit – An rnnlm implementation for training on huge datasets and very large vocabularies and usage in real-world ASR and MT problems
• GNU Gneural Network – GNU package which implements a programmable neural network
• IDLF – Intel® Deep Learning Framework; supports OpenCL
• Keras – Deep Learning library for Theano and TensorFlow
• Lasagne – Lightweight library to build and train neural networks in Theano
• Leaf – "The Hacker's Machine Learning Engine"; supports OpenCL
• MatConvNet – CNNs for MATLAB
• MXNet – Lightweight library for Python, R, Julia, Go, Javascript and more
• neon – Nervana's Python based Deep Learning framework
• Pylearn2 – Machine learning library mainly built on top of Theano
• scikit-neuralnetwork – Multi-layer perceptrons as a wrapper for Pylearn2
• Theano-Lights – Deep learning research framework based on Theano
• Veles – Distributed machine learning platform by Samsung

Related software

• Numenta Platform for Intelligent Computing, Numenta's open source implementation of their hierarchical temporal memory model

See also

• Deep learning#Software libraries
• List of datasets for machine learning research
• Comparison of datasets in machine learning
• Comparison of numerical analysis software
• Comparison of statistical packages
• List of numerical analysis software

References

1. http://caffe.berkeleyvision.org/installation.html
2. https://github.com/sh1r0/caffe-android-lib
3. https://github.com/MSRDL/caffe
4. https://github.com/niuzhiheng/caffe
5. http://caffe.berkeleyvision.org/development.html
6. http://caffe.berkeleyvision.org/tutorial/interfaces.html
7. https://github.com/BVLC/caffe/tree/opencl
8. https://github.com/BVLC/caffe/pull/2195
9. https://github.com/amd/OpenCL-caffe
10. https://github.com/BVLC/caffe/wiki/Model-Zoo
11. https://github.com/BVLC/caffe/issues/1207
12. https://github.com/Microsoft/CNTK/blob/master/LICENSE.md
13. https://github.com/Microsoft/CNTK/wiki/Setup-CNTK-on-your-machine
14. https://github.com/Microsoft/CNTK/wiki/CNTK-usage-overview
15. https://github.com/Microsoft/CNTK/issues/175
16. https://github.com/Microsoft/CNTK/issues/59#issuecomment-178104505
17. http://www.cntk.ai/
18. https://github.com/Microsoft/CNTK/wiki/ConvertDBN-command
19. https://github.com/deeplearning4j/nd4j/issues/27
20. http://nd4j.org/gpu_native_backends.html
21. http://deeplearning4j.org/model-zoo
22. https://singa.incubator.apache.org/
23. https://blogs.technet.microsoft.com/inside_microsoft_research/2015/12/07/microsoft-computational-network-toolkit-offers-most-efficient-distributed-deep-learning-computational-performance/
24. http://systemml.apache.org/community-members.html
25. https://github.com/tensorflow/tensorflow/issues/17
26. https://github.com/tensorflow/tensorflow/issues/22
27. http://deeplearning.net/software/theano/tutorial/using_gpu.html
28. https://github.com/Lasagne/Recipes/tree/master/modelzoo
29. https://github.com/soumith/torch-android
30. https://github.com/jonathantompson/jtorch
31. https://github.com/torch/torch7/wiki/Cheatsheet#opencl
32. https://github.com/torch/torch7/wiki/Cheatsheet#cuda
33. https://github.com/torch/torch7/wiki/ModelZoo