Over the past month, MyBridge has carefully compared nearly 250 new open source machine learning projects and selected the 10 best ones based on their usefulness to developers.

  • These projects have an average star count of 792 on GitHub

  • Topics covered include: research frameworks, automated machine learning libraries, evolutionary algorithms toolbox, object detection, games, etc.

No.1

Swapping is a framework for rapidly prototyping enhanced learning algorithms. [5507 stars on GitHub]

Project address: github.com/google/dopa…

No.2

TransmogrifAI: An automated machine learning library for creating modular, reusable, strongly typed machine learning workflows on Spark with minimal manual debugging effort. [947 stars on GitHub]

Project address: github.com/salesforce/…

No.3

Deep – the Exemplar – -based – Colorization: Microsoft researchers presented a paper called Deep Exemplar-based Colorization, or “Samples-based Deep Learning Coloring,” at SIGGRAPH 2018 this year, which is the industry’s first samples-based local coloring method for Deep learning. This project is the source code of this paper. [96 stars on GitHub]

Project address: github.com/msracver/De…

No.4

YOLOv3: Trained and implemented YOLOv3, an advanced real-time object detection system, with PyTorch. [567 stars on GitHub]

Project address: github.com/ultralytics…

No.5

Mantra: A high-level, rapid development framework for machine learning projects. [GitHub 286 stars]

Project address: github.com/RJT1990/man…

No.6

FastTSNE: Implemented T-SNE, a machine learning algorithm for dimensionality reduction, in a fast and parallel way. The objective of this project is to quickly implement T-SNE in one place, without the need for external C/C ++ dependencies. [280 stars on GitHub]

Project address: github.com/pavlin-poli…

No.7

Evolute: A simple and practical tool for fast experimentation with numerically optimized evolutionary algorithms. [GitHub 118 stars]

Project address: github.com/csxeba/evol…

No.8

AIF360: An open source library from IBM for detecting and removing bias in machine learning models. Includes datasets & model metrics for testing biases, explanations of these metrics and algorithms for mitigating biases in datasets and models. [GitHub 208 stars]

Project address: github.com/IBM/AIF360

No.9

DeepSort: an AI image annotation tool based on DeepDetect. When you’re stuck with a pile of poorly named images and can’t see what you’ve saved, this AI tool helps you quickly sort and annotate them. [111 stars on GitHub]

Project address: github.com/CorentinB/D…

No.10

Zombie-shooter – neural-network: the source code of the small game “AI fight zombies”, which is a small game where a developer uses Neural networks and enhanced learning to train an AI, and finally the AI learns to fight zombies. 7 stars on GitHub

Project address: github.com/Daporan/Zom…

Past: Poke here