Selected from Mybridge
Heart of the machine compiles
Participation: Li Zannan
Which machine learning projects got the most attention in 2017? Mybridge has put together a list of the Top 30 for us, with GitHub links to all of the projects below.
We compared nearly 8,800 open machine learning projects and selected the best 30. This is a very competitive list of excellent machine learning libraries, datasets and applications open sourced from January to December 2017. Mybridge AI rates them by popularity, engagement, and freshness. To give you an idea: their GitHub average stars is 3,558.
Open source projects make a lot of sense for data scientists to read the source code and build on what has been done before. Now, feel free to try some of last year’s best projects.
No.1
FastText: FastText representation/categorization library, Facebook (GitHub 11,786 stars)
Link: https://github.com/facebookresearch/fastText
Facebook releases a new version of fastText: Mobile and tutorials
Ps Muse: Multilingual unsupervised/supervised Word embedding, FastText (GitHub 695 stars)
Link: https://github.com/facebookresearch/MUSE
No.2
Deep Photo-style Transfer: Code and Data from Fujun Luan’s Paper Deep Photo Style Transfer (GitHub 9747 stars)
Link: https://github.com/luanfujun/deep-photo-styletransfer
No.3
Face Recognition: The simplest Python command line face recognition API from Adam Geitgey (GitHub 8672 stars)
Link: https://github.com/ageitgey/face_recognition
Python based open source face recognition library: 99.38% offline recognition rate
No.4
Magenta: Machine Intelligence Music and Art Generator (GitHub 8113 stars)
Link: https://github.com/tensorflow/magenta
How does Google’s Magenta project teach neural networks to write music?
No.5
Sonnet: Neural network Library based on TensorFlow (GitHub 5731 stars), by DeepMind member Malcolm Reynolds
Link: https://github.com/deepmind/sonnet
Sonnet: Build neural networks quickly in TensorFlow
No.6
GitHub 5462 Stars deeplearn.js: Web hardware Accelerated Machine Learning library from Google Brain team Nikhil Thorat
Link: https://github.com/PAIR-code/deeplearnjs
Deeplearn.js: Hardware-accelerated machine learning for web pages
No.7
Fast Style Transfer: TensorFlow Fast Style Transfer by Logan Engstrom from MIT (GitHub 4843 Stars)
Link: https://github.com/lengstrom/fast-style-transfer
No.8
Pysc2: StarCraft 2 Learning Environment from DeepMind Timo Ewalds et al. (GitHub 3683 Stars)
Link: https://github.com/deepmind/pysc2
No.9
AirSim: Open Source Autonomous driving simulator based on Unreal Engine by Shital Shah et al. (GitHub 3861 stars)
Link: https://github.com/Microsoft/AirSim
No.10
Facets: Machine Learning Data Set Visualization Tool, From Google Brain (GitHub 3371 stars)
Link: https://github.com/PAIR-code/facets
Facets: Looking at data in a new Light
No.11
Style2Paints: AI Comics Line draft Coloring tools, from Suzhou University (GitHub 3310 stars)
Link: https://github.com/lllyasviel/style2paints
Style2paints: Professional AI cartoon line draft automatic coloring tool
No.12
Tensor2Tensor: Tools for generalized Sequence-sequence models Ryan Sepassi at Google Brain (GitHub 3087 Stars)
Link: https://github.com/tensorflow/tensor2tensor
Learn It all: Tensor2Tensor, Google’s open source modular deep learning system
No.13
CycleGAN and Pix2pix in PyTorch: An image-to-image conversion tool based on PyTorch, from UC Berkeley PhD Student, GitHub 2847 stars
Link: https://github.com/junyanz/pytorch-CycleGAN-and-pix2pix
You draw graffiti, artificial intelligence to generate “cats” : Edges2Cats image conversion details
No.14
Faiss: Tool Library for Efficient Similarity Search and clustering with Dense Vectors, Facebook (GitHub 2629 stars)
Link: https://github.com/facebookresearch/faiss
No.15
Fashion-mnist: A MnIST-like Fashion Product Dataset from Zalando Tech Han Xiao (GitHub 2780 stars)
Link: https://github.com/zalandoresearch/fashion-mnist
No. 16
ParlAI: A framework for training and evaluating AI models on a variety of publicly available conversational datasets, Alexander Miller, Facebook (GitHub 2578 Stars)
Link: https://github.com/facebookresearch/ParlAI
ParlAI: An open source Artificial intelligence framework from Facebook that makes it easy to train and evaluate conversation models
No.17
Fairseq: Sequence to Sequence Toolkit from FAIR (GitHub 2571 Stars)
Link: https://github.com/facebookresearch/fairseq
Facebook’s new CNN machine Translation: Nine times faster and more accurate than Google
No.18
Pyro: Deep General-purpose Probability Programming with Python and PyTorch, Uber AI Labs (GitHub 2387 Stars)
Link: https://github.com/uber/pyro
Uber and Stanford Open source Deep probability programming language Pyro: Based on PyTorch
No.19
IGAN: Interactive Image Generator based on GAN (GitHub 2369 Stars)
Link: https://github.com/junyanz/iGAN
Berkeley and iGAN, Adobe’s open source deep learning image editing tool
No.20
Dmitry Ulyanov (GitHub 2188 stars) from Skoltech: Deep image-prior: Image restoration using neural networks without learning process
Link: https://github.com/DmitryUlyanov/deep-image-prior
No.21
Face Classification: Real-time Face Detection and Expression/Gender Classification based on Keras CNN Model with OpenCV, Training with FER2013 / IMDB Dataset (GitHub 1967 Stars)
Link: https://github.com/oarriaga/face_classification
No.22
Speech to Text WaveNet: End-to-end Sentence Level English Speech Recognition using DeepMind’s WaveNet and TensorFlow, Namju Kim from Kakao Brain (GitHub 1961 Stars)
Link: https://github.com/buriburisuri/speech-to-text-wavenet
DeepMind WaveNet reduces the gap between machine-synthesized speech and human speech by 50%
No.23
StarGAN: Unified Generative Adversarial Networks for Multi-domain Image-image Transformations (GitHub 1954 Stars)
Link: https://github.com/yunjey/StarGAN
No.24
Mi-agents: Unity Machine Learning Agents, Arthur Juliani from Unity3D (GitHub 1658 stars)
Link: https://github.com/Unity-Technologies/ml-agents
No.25
Deep Video Analytics: Distributed Visual Search and Visual Data Analysis Platform, Akshay Bhat, Cornell University (GitHub 1494 Stars)
Link: https://github.com/AKSHAYUBHAT/DeepVideoAnalytics
No.26
OpenNMT: Open Source Neural Machine Translation on Torch (GitHub 1490 Stars)
Link: https://github.com/OpenNMT/OpenNMT
Harvard NLP Group open Source Neural machine translation toolkit OpenNMT: ready for production
No.27
Pix2PixHD: Image synthesis and processing with conditional GAN at 2048×1024 resolution, from Nvidia AI Scientist Ming-Yu Liu (GitHub 1283 stars)
Link: https://github.com/NVIDIA/pix2pixHD
No.28
Horovod: Distributed TensorFlow Training Framework from Uber Engineering Team (GitHub 1188 stars)
Link: https://github.com/uber/horovod
Horovod: Uber’s open-source TensorFlow distributed deep learning framework
No.29
Ai-blocks: Powerful and intuitive WYSIWYG interface that allows anyone to create machine learning models (GitHub 899 Stars)
Link: https://github.com/MrNothing/AI-Blocks
No.30
Dabi Ahn, Kakao Brain Team (GitHub 845 stars) Voice Conversion with Non-Parallel Data: Deep Neural Network Speech Style Conversion based on TensorFlow
Link: https://github.com/andabi/deep-voice-conversion
The original link: https://medium.mybridge.co/30-amazing-machine-learning-projects-for-the-past-year-v-2018-b853b8621ac7
This article is compiled for machine heart, reprint please contact this public number for authorization.
✄ — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — —
Join Heart of the Machine (full-time reporter/intern) : [email protected]
Contribute or seek coverage: [email protected]
Advertising & Business partnerships: [email protected]