GitHub is one of the most popular sources and this year GitHub featured a lot of open source projects. It also saw a record number of new users coming to GitHub and hosted over 100 million repositories. While there have been a lot of projects, there were a few that grabbed more popularity than the others. In this article, we list the top 10 open source projects by unique contributors that were used the most, which is largely decided by the number of stars received by the projects.
Here are the top 10 open source projects in Machine Learning that were popular in 2018
1| BERT
BERT or Bidirectional Encoder Representations from Transformers is an all-new method of pre-training language representations. It is the first unsupervised, deeply bidirectional system for pre-training natural language processing (NLP) and obtains new state-of-the-art results on eleven NLP tasks. This repository contains TensorFlow code and pre-trained models for BERT.
2| DeepCreamPy
DeepCreamPy is a deep learning based tool that is used to automatically replace censored artwork in hentai with probable reconstructions. The features contain are higher quality decensoring images of any size and shape, mosaic decensor support and user interface (WIP).
3| Horizon
This is an open source end-to-end platform for Applied Reinforcement Learning (Applied RL), built in Python that uses PyTorch for modelling and training as well as Caffe2 for model serving. It is mainly used in Facebook and algorithms like Soft Actor-Critic (SAC), DDPG, DQN are supported here.
4| TRFL
Pronounced as truffle, this is a library built on top of TensorFlow and is useful for building blocks for writing reinforcement learning (RL) agents in both CPU and GPU versions of TensorFlow.
5| DeOldify
DeOldify by Jason Antic, the name says it all. It is a deep learning based project that is used to colorize and restoring the old black and white images into a colourful one.
6| AdaNet
It is basically a lightweight TensorFlow based network and builds on AutoML efforts that is used for automatically learning high-quality models with the least expert interference. The goals concerned in this project are easy usability, flexibility, speed and guarantee of learning.
7| Graph Nets
The working of Graph Nets is that it takes graph as input and returns graph as output. It also validates the deep learning architecture for learning and understanding the rules, relations and entities in a graph. Google-owned and London-headquartered DeepMind, opened the graph nets library in October. It can be installed and used in TensorFlow.
8| The MAME RL Algorithm Training Toolkit
This Python Library is used on almost any arcade games to train a reinforcement learning algorithm. Based on the Linux operating system with version 3.6+, it allows your algorithm to step through gameplay while receiving and sending actions to interact with the game.
9| PocketFlow
It is an open source framework that is used for compressing and accelerating deep learning models. The main goal is to provide an easy and usable toolkit to improve efficiency for developers with minimal degradation of performance.
10| Maskrcnn-Benchmark (Faster R-CNN and Mask R-CNN in PyTorch 1.0)
Released under MIT license, this project mainly uses PyTorch 1.0 and aims at bestowing the requisite building blocks for creating detection and segmentation models without any difficulty.
