Recommend 10 open source AI projects that are highly praised and have unique features

About artificial intelligence projects, I believe that we have seen or used a lot, but most of them look very “high”, let a person feel to master them like learning the dragon slaying skills. In fact, there are many ai projects that are useful and interesting to use. Here are 10 open source AI projects with unique features.

STYLE2PAINTS: Powerful AI for paints

https://www.oschina.net/p/style2paints

Please enter the title bcdef

Recommended reason: The new generation of powerful line draft color AI, can be uploaded according to the user’s custom color line draft color. The project provides online access to the site, which is very easy to use.

Please enter the title abcdefg

SerpentAI: A learning framework for teaching AI to play games

https://www.oschina.net/p/serpentai

Please enter the title bcdef

Why you should check it out: SerpentAI aims to provide a valuable tool for machine learning and AI research. But at the same time, it is also very interesting for lovers.

Please enter the title abcdefg

Synaptic. Js: Neural network library for browsers

https://www.oschina.net/p/synapticjs

Please enter the title bcdef

Why you should check it out: Synaptic. Js is a JavaScript neural network library for Node.js and browsers that can build and train basically any type of first or even second order neural network.

Four classical neural network algorithms are built into the project: Multilayer perceptrons, Multilayer long-short Term memory networks, Liquid State machines Machine), Hopfield neural network. With Synaptic. Js, you can easily test and compare the performance of different architectures.

Please enter the title abcdefg

Snake-AI: Artificial intelligence for Snake games

https://www.oschina.net/p/snake-ai

Please enter the title bcdef

Why you should check it out: An AI for snake game written in C/C++. Using shortest path, longest path, artificial intelligence algorithm.

The AI’s goal is to get the snake to eat as much food as possible until the entire map is full.

Demo

Please enter the title abcdefg

Uncaptcha

https://www.oschina.net/p/uncaptcha

Please enter the title bcdef

Why we recommend it: Cracking the AI algorithm for reCAPTCHA systems. The unCAPTCHA algorithm beat the Google reCAPTCHA system 85% of the time. It relies on audio captcha attacks – using browser automation software to parse the necessary elements and recognize speech numbers, and deliver those numbers programmatically, ultimately successfully tricking the target website.

Please enter the title abcdefg

Sockeye: Neural machine translation framework

https://www.oschina.net/p/sockeye

Please enter the title bcdef

Why you should check it out: Sockeye is a fast and extensible deep learning library based on Apache MXNet.

The Sockeye codebase has a unique advantage from MXNet. For example, Sockeye combines declarative and imperative programming styles through symbolic and imperative MXNet apis; It can also train models on multiple Gpus in parallel.

Sockeye implements the current best sequence-to-sequence model on MXNet. It also provides appropriate defaults for all sequence-to-sequence model hyperparameters. For optimization, there is no need to worry about stopping criteria, metric tracking, or weight initialization. You can simply run the provided training command line interface (CLI) and easily change the underlying model architecture.

Please enter the title abcdefg

Php-ml: PHP machine learning library

https://www.oschina.net/p/php-ml

Please enter the title bcdef

We all know that Python or C++ offer more machine learning libraries, but most of them are complex enough to make configuration a pain for beginners.

Php-ml machine learning library although there is no particularly sophisticated algorithms, but it has the most basic machine learning, classification and other algorithms, small projects or small companies to do some simple data analysis, prediction and so on enough.

Php-ml is a machine learning library written in PHP. It also includes algorithm, cross validation, neural network, preprocessing, feature extraction and so on.

Please enter the title abcdefg

CycleGAN: Generates adversarial network image processing tools

https://www.oschina.net/p/cyclegan

Please enter the title bcdef

Why you should check it out: It’s a powerful tool that not only “restores” a painting to a photo (sort of a “backfilter”), but also turns summer into winter or a regular horse into a zebra.

Unlike other AI paintings, CycleGAN’s team tried to build a two-way algorithm that could be converted in both directions without losing information.

In CycleGAN, the details of the photos are required to be completely preserved, and the researchers hope to be able to input an image into CycleGAN and transform it many times (photo → painting → photo → painting → photo), and eventually get the same or similar image as the original one.

Please enter the title abcdefg

Deeplearn.js: a machine learning JS library for accelerating hardware

https://www.oschina.net/p/deeplearn-js

Please enter the title bcdef

Deeplearn.js is an open source JavaScript library from Google that can be used for machine intelligence and accelerate WebGL. Deeplearn.js runs completely in the browser, requires no installation, and requires no back-end processing.

Deeplearn.js provides efficient machine learning building blocks that allow us to train neural networks in a browser or run pre-trained models in inferential mode. It provides an API for building a differentiable data flow graph, as well as a series of straightforward mathematical functions.

While machine learning libraries on browsers have existed for years (e.g. Andrej Karpathy’s ConvnetJS), they are limited by JavaScript speed or are limited to reasoning and cannot be used for training (e.g. TensorFire).

In contrast, deeplearn.js achieves significant acceleration by taking advantage of WebGL’s ability to perform computations on gpus and perform full backpropagation.

Please enter the title abcdefg

TensorFire: browser-side neural network framework

https://www.oschina.net/p/tensorfire

Please enter the title bcdef

Why you should check it out: TensorFire is a WebGL-based neural network framework that runs in a browser. Applications written using TensorFire can implement cutting-edge deep learning algorithms while running directly in modern browsers without any installation or configuration.

TensorFire is nearly a hundredfold faster than previous neural network frameworks in some browsers, even matching the performance of code running on native cpus.

Developers can also use the underlying interfaces provided by TensorFire for other high-performance computing, such as PageRank, cellular automata simulation, image conversion and filtering, and more.

There are other great open source AI projects that have not yet appeared in this article, so please leave your recommendations in the comments

SDCC 2017 Database Online Summit is coming soon. Adhering to the content principle of dry goods and solid materials (case), we invited database experts from Alibaba, Tencent, Weibo, netease and other enterprises and university researchers. Focus on hot database technologies such as Oracle, MySQL, PostgreSQL, Redis, etc., from the deep digging of core technologies to the analysis of high availability practices, create the essence of compressed sharing, draw infersions from one point of view, think and fight, please click “read the original article” or scan the following TWO-DIMENSIONAL code to see more details.