In 2017, we compared nearly 8,800 open source machine learning projects and selected the top 30 (0.3% chance).

This is a very competitive ranking, and I have handpicked the best open source machine learning libraries, datasets and applications released between January and December 2017. Mybridge AI assesses the quality of these entries by taking into account metrics such as popularity, engagement and recency. To give you an idea of the quality, the average number of GitHub stars is 3,558.

Open source projects are very useful for data scientists. You can learn by reading the source code and then build your own on top of existing projects. Now let’s get easy on the machine learning projects you might have missed over the past year.

List of recommended study books:

A) Neural networks

 

Deep Learning A-Z: Hands-on Artificial Neural Networks

[68,745 recommendations, 4.5/5 stars]

B） TensorFlow

 

The Complete Guide to TensorFlow for Deep Learning with Python

[17,834 recommendations, 4.6/5 stars]

 

Open Source Project Leaderboards

First place

FastText: FastText representation/classification library.

[11,786 stars on Github] courtesy of Facebook Research

 

[Muse: Multilingual unsupervised or supervised word embedding based on fast Text. 695 stars on Github]

second

Deep-photo-style Transfer: Code and data for the paper “Deep Photo Style Transfer”.

[9747 stars on Github] courtesy of Dr. Fujun Luan, Cornell University

 

Third,

Face Recognition: The simplest Python command-line face recognition API.

[8672 stars on Github] courtesy of Adam Geitgey

 

4,

Magenta: Machine Intelligent music and art generator.

[There are 8,113 stars on Github]

 

fifth

Sonnet: A neural network library based on TensorFlow.

[5731 stars on Github]. Courtesy of Deepmind’s Malcolm Reynolds

 

6

Deeplearn.js: a hardware-accelerated machine intelligence library for the Web.

[5,462 stars on Github]. Courtesy Nikhil Thorat of Google Brain

 

seventh

Fast Style Transfer: Fast Style Transfer for TensorFlow.

[4843 stars on Github] courtesy of Logan Engstrom, MIT

 

eighth

Pysc2: StarCraft II learning environment.

[3,683 stars on Github] courtesy of DeepMind’s Timo Ewalds

 

The ninth

AirSim: open source simulator based on Unreal Engine for Microsoft AI&Research autonomous vehicles.

[3,861 stars on Github]. Provided by Shital Shah of Microsoft

 

10th

Facets: Machine learning data set visualization tool.

[3371 stars on Github] courtesy of Google Brain

 

11th

Style2Paints: AI image colorization.

[There are 3,310 stars on Github]

 

12th

Tensor2Tensor: A library of tools for generalized sequence-sequence models

[3087 stars on Github] courtesy of Ryan Sepassi of Google Brain

 

13th

CycleGAN and Pix2pix in PyTorch: Image-to-image conversion tools based on PyTorch (e.g. Horse2zebra, Edges2Cats, etc.).

[2847 stars on Github] contributed by Dr. Jun-yan Zhu, Berkeley

 

14th

Faiss: library for efficient similarity search and clustering of dense vectors.

[2,629 stars on Github]. Courtesy of Facebook Research

 

15

Fashion-mnist: A Fashion product data set similar to MNIST.

[2,780 stars on Github] courtesy of Zalando Tech research scientist Han Xiao

 

16

ParlAI: A framework for training and evaluating artificial intelligence models on a variety of publicly available conversational data sets.

[2,578 stars on Github]. Here’s Alexander Miller from Facebook Research

 

17

Fairseq: Facebook AI Research Sequence – a toolkit for sequences.

[There are 2,571 stars on Github]

 

18th

Pyro: Deep general-purpose probabilistic programming using Python and PyTorch.

[2387 stars on Github] courtesy of Uber Engineering

 

19 in

IGAN: Interactive image generator supported by GAN.

[There are 2,369 stars on Github]

 

20

Deep-image-prior: the neural network is used for image restoration without learning the neural network.

[2188 stars on Github] by Dmitry Ulyanov, Ph.D. Skoltech

 

21,

Face_classification: Real-time face detection and emotion/gender classification using fer2013/ IMDB datasets with KerAS CNN model and openCV.

[1967 star on Github]

 

22nd

Speech-to-text-wavenet: End-to-end sentence-level English Speech recognition using DeepMind’s WaveNet and TensorFlow.

[1961 star on Github] courtesy of Namju Kim of Kakao Brain

 

23

StarGAN: Unified generative adversarial network for multi-domain image-image transformations.

[1954 stars on Github] by Yunjey Choi, Korea University

 

24th

Ml-agents: Unity machine learning agent.

Deep Learning at Unity3D by Unity Juliani

 

25th

DeepVideoAnalytics: Distributed visual search and visual data analysis platform.

[1494 stars on Github] by Akshay Bhat, PhD, Cornell University

 

26

OpenNMT: Open source neural machine translation on Torch.

[There are 1,490 stars on Github]

 

27,

Pix2pixHD: Use conditional GANs to synthesize and process 2048×1024 resolution images.

[1283 stars on Github] courtesy of Nvidia AI Research scientist Ming-Yu Liu

 

28th

Horovod: Distributed training framework for TensorFlow.

[1188 stars on Github] courtesy of Uber Engineering

 

29

Ai-blocks: A powerful and intuitive WYSIWYG interface that allows anyone to create machine learning models.

[899 stars on Github]

 

30th

Voice Conversion with non-Parallel Data: Deep neural network speech Conversion (or speech style Conversion) based on Tensorflow.

[845 stars on Github] contributed by Dabi Ahn of AI Kakao Brain

 

 

This article is recommended by Beijing Post @ Love coco – Love life teacher, translated by Ali Yunqi Community organization.

30 Amazing Machine Learning Projects for the Past Year (V. 2018)

Author: Mybridge

Translator: Altman, edited by Yuan Hu.

The article is a brief translation. For more details, please refer to the original text