Deep learning is a machine learning method based on the demonstration of data. It has been developed and popular in recent years.
As a relatively new concept, for beginners who want to enter the field, or for those who are already familiar with the method, the learning resources at your fingertips are too rich.
In order not to be eliminated by the ever-changing technology and trends, it is a good way to actively participate in the learning and interaction of open source projects in the deep learning community.
In this article, the Chinese editor will introduce you to the 16 most popular deep learning open source platforms and open source libraries in GitHub. In addition, some of the better platforms and frameworks have not entered the list, and Wen Xiaobian also listed Come out for your reference.
GitHub has the highest collection of 16 open source deep learning frameworks. The greener the circle, the newer the frame, and the more blue the color indicates the earlier the frame.
As you can see from the above picture, TensorFlow tops the list, and the second and third place are Keras and Caffe. The following small series will share these resources with everyone.
16 of the best deep learning open source frameworks and platforms
1. TensorFlow
TensorFlow was originally developed by researchers and engineers at Google Brain Team in Google's Machine Intelligence research organizaTIon. This framework is designed to facilitate researchers' research on machine learning and to simplify the process of migration from research models to actual production.
Collection: 96655, Contributions: 1432, Number of program submissions: 31714, Creation date: November 1, 2015.
Link: https://github.com/tensorflow/tensorflow
2 Keras
Keras is an advanced neural network API written in Python that works with TensorFlow, CNTK or Theano.
Collection: 28385, Contributions: 653, Number of submissions: 4468, Creation date: March 22, 2015.
Link: https://github.com/keras-team/keras
3. Caffe
Caffe is a deep learning framework focused on expressiveness, speed and modularity, developed by the Berkeley Vision and Learning Center and community contributors.
Collection: 23750, Contributions: 267, Number of program submissions: 4128, Creation date: September 8, 2015.
Link: https://github.com/BVLC/caffe
4. Microsoft CogniTIve Toolkit
The Microsoft CogniTIve Toolkit (formerly CNTK) is a unified set of deep learning tools that describe a neural network as a series of computational steps represented by directed graphs.
Collection: 14243, Contributors: 174, Number of program submissions: 15613, Creation date: July 27, 2014.
Link: https://github.com/Microsoft/CNTK
5. PyTorch
PyTorch is a framework for tensor computing and dynamic neural networks with powerful GPU support combined with Python.
Collection: 14101, Contributions: 601, Number of program submissions: 10733, Creation date: January 22, 2012.
Link: https://github.com/pytorch/pytorch
6. Apache MXnet
Apache MXnet is a deep learning framework designed to increase efficiency and flexibility. It allows users to mix symbolic programming with imperative programming to maximize efficiency and productivity.
Collection: 13699, Contributions: 516, Number of program submissions: 6953, Creation date: April 26, 2015.
Link: https://github.com/apache/incubator-mxnet
7. DeepLearning4J
DeepLearning4J, like ND4J, DataVec, Arbiter and RL4J, is part of the Skymind Intelligence Layer. It is an open source distributed neural network library written in Java and Scala and is certified by Apache 2.0.
Collection: 8725, Contributions: 141, Number of program submissions: 9647, Creation date: November 24, 2013.
Link: https://github.com/deeplearning4j/deeplearning4j
8. Theano
Theano handles user-defined, optimized, and computational mathematical expressions about multidimensional arrays efficiently. But in September 2017, Theano announced that there will be no further significant progress after the 1.0 release. But don't be disappointed, Theano is still a very powerful library to support your research in deep learning.
Collection: 8141, Contributions: 329, Number of program submissions: 27974, Creation date: January 6, 2008.
Link: https://github.com/Theano/Theano
9. TFLearn
TFLearn is a modular and transparent deep learning library built on top of TensorFlow to provide a higher level of API for TensorFlow to facilitate and speed up experimental research while maintaining full transparency and compatibility.
Collection: 7933, Contributions: 111, Number of program submissions: 589, Creation date: March 27, 2016.
Link: https://github.com/tflearn/tflearn
10. Torch
Torch is the main software package in Torch7, which defines data structures and mathematical operations for multidimensional tensors. In addition, it provides a number of utility software for accessing files, serializing any type of object, and more.
Collection: 7834, Contributions: 133, Number of program submissions: 1335, Creation date: January 22, 2012.
Link: https://github.com/torch/torch7
11. Caffe2
Caffe2 is a lightweight, deep learning framework with modularity and scalability. It has been improved on the basis of the original Caffe, which improves its expressiveness, speed and modularity.
Collection: 7813, Contributions: 187, Number of program submissions: 3678, Creation date: January 21, 2015.
Link: https://github.com/caffe2/caffe2
12. PaddlePaddle
PaddlePaddle (Parallel Distributed Deep Learning) is an easy-to-use, efficient, flexible, and scalable deep learning platform. Originally developed by Baidu scientists and engineers, it is designed to apply deep learning to Baidu's many products.
Collection: 6726, Contributors: 120, Number of program submissions: 13733, Creation date: August 28, 2016.
Link: https://github.com/PaddlePaddle/Paddle
13. DLib
DLib is a modern C++ toolkit that includes machine learning algorithms and tools for developing complex software based on C++ to solve real-world problems.
Collection: 4676, Contributions: 107, Number of submissions: 7276, Date of establishment: April 27, 2008.
Link: https://github.com/davisking/dlib
14. Chainer
Chainer is a stand-alone open source framework based on Python for deep learning models. It provides flexible, intuitive, high-performance means to implement a comprehensive deep learning model, including the latest recurrent neural networks and variational automata. Encoder (variaTIonal auto-encoders).
Collection: 3685, Contributions: 160, Number of program submissions: 13700, Creation date: April 12, 2015.
Link: https://github.com/chainer/chainer
15. Neon
Neon is a Python-based deep learning library developed by Nervana. It's easy to use and its performance is at its highest level.
Collection: 3466, Contributions: 77, Number of program submissions: 1112, Creation date: May 3, 2015.
Link: https://github.com/NervanaSystems/neon
16. Lasagne
Lasagne is a lightweight library that can be used to build and train neural networks on Theano.
Collection: 3417, Contributions: 64, Number of program submissions: 1150, Creation date: September 7, 2014.
Link: https://github.com/Lasagne/Lasagne
Draw-wire sensors of the wire sensor series measure with high linearity across the entire measuring range and are used for distance and position measurements of 100mm up to 20,000mm. Draw-wire sensors from LANDER are ideal for integration and subsequent assembly in serial OEM applications, e.g., in medical devices, lifts, conveyors and automotive engineering.
Linear Encoder,Digital Linear Encoder,Draw Wire Sensor,1500Mm Linear Encoder
Jilin Lander Intelligent Technology Co., Ltd , https://www.jilinlandermotor.com