GitHub recent 10 Most interesting machine learning open source project

Source: paperweekly

This article a total of 900 characters, recommended to read 6 minutes.
This article lists the top ten interesting machine learning open source projects for you recently GitHub.
-01-

Face recognition

#世界上最简单的人脸识别库

This project is known as the simplest face recognition library in the world and can be invoked using Python and the command line. The library is built using the Dlib-top depth learning face recognition technology, with an accuracy rate of 99.38% on the outdoor Face Detection database (labeled Faces in the Wild benchmark).

Project Link:

Https://github.com/ageitgey/face_recognition

-02-

MUSE

#多语言词向量 Python Library

The multilingual word vector Python Library, which is open source by Facebook, provides a multilingual word vector based on Fasttext implementations and a large, high-quality bilingual dictionary, including unsupervised and supervised. There are supervised methods using bilingual dictionaries or the same string, unsupervised methods that do not use any parallel data.

No supervision method can refer to Word translation without Parallel Data this paper.

Thesis Link:

https://www.paperweekly.site/papers/1097

Project Link:

Https://github.com/facebookresearch/MUSE

-03-

Foolnltk

#中文处理工具包

Features of this project:

• May not be the fastest open source Chinese participle, but it is probably the most accurate open source Chinese participle

• Based on bilstm model training

• including participle, pos tagging, entity recognition, there are relatively high accuracy rate

• User-defined dictionaries

Project Link:

Https://github.com/rockyzhengwu/FoolNLTK

-04-

Arnold

#最擅长玩 "The Destroyer" game AI

This project, from Carnegie Mellon University, is the Pytorch open source code for the 2017 Vizdoom "destroyer of Warriors" AI Death Championship champion Arnold.

Thesis Link:

https://www.paperweekly.site/papers/1440

Project Link:

Https://github.com/glample/Arnold

-05-

Bottom-up Attention VQA

#2017 VQA Challenge First place

This project is pytorch reappearance of two papers by 2017 VQA Challenge first team.

Papers: bottom-up and Top-down Attention for Image captioning and Visual question answering

Link: https://www.paperweekly.site/papers/754

Paper: Tips and Tricks for Visual question Answering:learnings from the 2017 Challenge

Link: https://www.paperweekly.site/papers/1441

Project Link:

Https://github.com/hengyuan-hu/bottom-up-attention-vqa

-06-

Yolov2-pytorch

#PyTorch Edition YOLOv2

The Pytorch version of the famous object detection library YOLOv2, this project can also convert the training model to fit Caffe 2.

Project Link:

Https://github.com/ruiminshen/yolo2-pytorch

-07-

Simple Railway Captcha Solver

#基于 CNN identification of ticket booking verification code for railway station

This project uses the simple convolutional neural network to make the verification code that identifies the railway booking website, the part of the training set is generated by imitating the style of the verification code, the part of the verification set is retrieved from the Taiwan Railway booking website, and then the manual method marks about 1000 pens.

At present, the single code identification rate of the verification set for the 6-code type is 98.84%, and the success rate of the whole identification is up to 91.13%.

Project Link:

Https://github.com/JasonLiTW/simple-railway-captcha-solver

-08-

Alphazero-gomoku

#用 Alphazero under Gobang

This is an implementation of the Alphazero algorithm in Gobang, because Gobang is much simpler than Weiqi or chess, so it takes only a few hours to train a good AI model.

Thesis: Alphazero:mastering Chess and shogi by Self-play with a general reinforcement Learning algorithm

Link: https://www.paperweekly.site/papers/1297

Thesis: Alphago zero:mastering the game of Go without human knowledge

Link: https://www.paperweekly.site/papers/942

Project Link:

Https://github.com/junxiaosong/AlphaZero_Gomoku

-09-

Gym-extensions

#OpenAI Gym Extension Set

This is a OpenAI Gym Library expansion package, which includes: multitask Learning, migration learning, inverse reinforcement learning functions.

Project Link:

Https://github.com/Breakend/gym-extensions

-10-

Myia

#Python Depth Learning Framework

Myia is a new Python depth learning framework with simple, automated differentiation, and performance optimization features.

Project Link:

Https://github.com/mila-udem/myia