September 2018 open source machine learning Github project TOP 10

Source: ATYUN AI platform

In the past month, we learn open source projects rank nearly 250 machine, select the top 10.

In the meantime, we will project compared with new or major release. Mybridge AI project are ranked according to a variety of factors to measure the quality of professionals.

The average rating for all projects on Github Star 728
Topics include: research framework, AutoML libraries, deep learning, PyTorch, TSNE, algorithm toolbox, Fairness-ai, Deepdetect, ZOMBIES

Open source projects useful for programmers. I hope you find a fun project to bring inspiration to you.

1.Dopamine

Strengthen the framework for rapid prototyping learning algorithm, which is designed to satisfy a small, easy to understand the needs of the code base, users are free to experiment with the idea of (speculative research). Design principles are: simple experiment: to let new users to easily run the benchmark test; flexible development: to let new users to easily try to study creative; compact and reliable: provides embodiments for a few algorithms battle-tested; repeat: to promote the results can be repeatability. It is provided by Google.

Github comment Star 5316.

Project: github.com/google/dopamine utm_source = mybridge & utm_medium = blog & utm_campaign = read_more?

2.TransmogrifAI

AutoML a library for the Spark build modular, its focus is on the development of machine learning machine learning automation accelerates the work of developers, as well as the implementation of compile-time type safety, modular and reusable API. By automation, it can achieve the accuracy of the model close to the manual adjustment, it can be reduced almost 100 times longer. Provided by Salesforce.

Github assessment star 902.

Project: github.com/salesforce/TransmogrifAI utm_source = mybridge & utm_medium = blog & utm_campaign = read_more?

3.Deep Exemplar-based Colorization

Deep Exemplar-based Colorization is a first learning method based on the local depth of the coloring of a sample. Given reference color image, a convolutional neural network mapping grayscale image directly to the output color image. The proposed network comprises two subnetworks, similar subnets, calculating semantic similarities between the reference and the target; coloring subnet and chrominance channels, their choice, and the propagation of the prediction target. By MSRA CVer.

Github assessment star 82.

Project: github.com/msracver/Deep-Exemplar-based-Colorization utm_source = mybridge & utm_medium = blog & utm_campaign = read_more?

4.YOLOv3

PyTorch in training and reasoning. The most advanced real-time object detection system. By Ultralytics.

Github assessment star 353.

Project:? Github.com/ultralytics/yolov3 utm_source = mybridge & utm_medium = blog & utm_campaign = read_more

5.Mantra

Rapid development framework is an advanced machine-learning projects. For writing the data set and model class, supports single training, with tracking the UI, and can be easily packaged and share results. By Ross Taylor.

main feature:

Common data set and the model type of Class Boilerplate
Command line interface parameters for resolution training
Automatically configure cloud instances for remote training
UI monitor for training, comparative experiments and storage media
By designing the package model and the data set, for easy sharing

Github assessment star 260.

Project: github.com/RJT1990/mantra utm_source = mybridge & utm_medium = blog & utm_campaign = read_more?

6.FastTSNE

tSNE fast parallel implementation. Interpolation using an FFT acceleration, calculated visualized within 2 minutes 160 796 single cells from mouse transcriptome of the nervous system. By PavlinPoličar.

Github assessment star 255.

Project: github.com/pavlin-policar/fastTSNE utm_source = mybridge & utm_medium = blog & utm_campaign = read_more?

7.Evolute

A simple tool to quickly test the numerical optimization of evolutionary algorithm used. It defines the group of individuals, it expressed as a floating point vector, and the predefined set of sequence evolution configurable operator applied to them. By CsabaGór.

Github assessment star 111.

Project: github.com/csxeba/evolute utm_source = mybridge & utm_medium = blog & utm_campaign = read_more?

8.AIF360

An open source library can help detect and eliminate bias machine learning model. This includes a set of metric data model and test the deviation, and the explanation of these metrics for mitigating algorithm deviation data set and the model. Provided by IBM.

Github assessment star 159.

Project:? Github.com/IBM/AIF360 utm_source = mybridge & utm_medium = blog & utm_campaign = read_more