Look at the data. What scientists are using: ten deep learning projects on GitHub _deeplearning

The author Matthew May is a computer postgraduate in parallel machine learning algorithms, and Matthew is also a data mining learner, a data enthusiast, and a dedicated machine-learning scientist. Open source tools play an increasingly important role in data science workflows. GitHub Ten deep learning projects, which include a large number of code base, framework and learning materials. Look at what tools people on GitHub are using and what resources they are learning from.

The latest Kdnuggets software survey found that over the past 12 months, 73% of data scientists have used free data science tools. It goes without saying that Python and R languages (both open source) and their ecosystems are the most prominent and indispensable data scientists use in these free data science tools.

GitHub has become a veritable Open-source software Exchange Center, hosting almost any type of project it wants. Deep learning is becoming more and more popular in academic, research and interest, and is becoming more and more important in data science, and we will explore the github of deep learning projects in the field of excellence.

It should be noted that most of the important projects that are considered to be of deep learning do not appear on the list because they are not involved in GitHub search for "deep learning".

1. Caffe

Caffe is a library of deep learning programs created with Python and Matlab. Founded by the Berkeley Vision and Learning Center at Berkeley, it is understandable that people can only use Caffe as a computer vision application; In fact, this is a general-purpose deep learning library that enables the development of convolution networks and the creation of other applications such as vision and voice.

2. Data Science IPython Notebooks

This is the Ipython notebook that Donne Martin plans to collect. Topics cover Big Data, Hadoop, Scikit-learn and science python stacks, and many other things. As for deep learning, frameworks such as TensorFlow, Theano and Caffe are also covered, and of course there are specific structures and concepts involved.

3. Convnetjs

Written by Stanford PhD Andrej Karpathy, he has also maintained a blog update. Convnetjs is a neural network implemented by JavaScript and its common modules, and includes a large number of browser-based instances. These documents and instances are numerous and complete. Don't let JavaScript and neural networks combine to scare you away, which is a very popular and useful project.

4. Keras

Keras is also a library of Python deep learning programs, but it leverages TensorFlow and Theano, which means it can run on either of the 2 most popular in-depth learning and development libraries currently known. It is also one of the more and more highly described libraries, all of which are very similar: abstracting the underlying deep learning engine, allowing users to deploy neural networks faster, more easily, and more flexibly. Keras supports the mainstream depth learning architecture, with a 30-second QuickStart Guide, and complete documentation.

5. Mxnet

As a framework for deep learning, mxnet is designed to be flexible and efficient, and to improve code productivity by allowing a mix of instruction programming and symbolic programming techniques. This project can be bundled with multiple languages, such as Python, R, and Julia. Mxnet also comes with a series of neural network guides and blueprints. It is also noteworthy that a related project uses JavaScript to implement mxnet in a browser environment where interested friends can test a graphics classification model.

6. Qix

This is a library of GitHub versions of various computing and programming topics related to resources, including Node.js, Golang, and depth learning. The reason for this "seems to be (appears)" Is that the GitHub version of the library is written in Chinese, and the translations offered by Google are even more puzzling. However, there are many links, so if you can speak Chinese or understand Chinese, perhaps there are some valuable things here.

7. deeplearning4j

DEEPLEARNING4J is an industry-intensive, in-depth learning framework for Java and Scala. As the only one of the JVM's deep learning solutions worth studying, it has an obvious advantage in this field. Not only is it a good combination of Hadoop and spark, it can also use a GPU. His documents and guidelines are also very reliable.

8. Machine Learning Tutorials

This is a list of machine learning and depth learning tutorials, articles and resources. This list is organized by topic and includes a number of categories related to deep learning, including computer vision, enhanced learning, and various architectures. You can also click here to make some contributions because of the wide range of content that has been known in social media for months.

9. Deeplearntoolbox

Deepleantoolbox is a deep learning toolkit for use in MATLAB and octave. Unfortunately, this project has now been scrapped and stopped maintenance. The GitHub version library also points to other valuable choices in the depth of learning: Theano and TensorFlow.

If there is any value in this abandoned cloud version library, this is the link, and this tutorial, written by Yoshua Bengio, is included in this repository as a learning resource for learning the depth of the learning architecture used by AI.

LISA Lab Deep Learning Tutorials

This GitHub version library summarizes the practice materials for the deep study Course at the University of Montreal, Canada. The following documents are extracted:

This practice kit will introduce you to some of the most important depth learning algorithms and show how to use Theano to run these algorithms.

Theano is a python library that makes writing depth learning models simpler and allows users to select the GPU to train them.

Click here to view the original link of the course materials

This article permanently updates the link address: http://www.linuxidc.com/Linux/2016-01/127621.htm