在过去的一年里,Mybridge AI 比较了近 8,800个开源机器学习项目,选择了前30名(概率只有0.3%)。
这是一个竞争激烈的名单,精挑细选了2017年1月到12月之间发布的最佳开源机器学习库、数据集和应用程序。Mybridge AI 通过考量受欢迎程度、参与度和新近度来等指标来评估这些参选项目的质量。这些项目在Github得星的数量平均为3,558个,其质量相当不错。
开源项目对数据科学家来说可能大有裨益。 通过阅读源代码并在现有项目之上构建一些东西来可以加快学习步伐。是该拿出时间来玩玩过去一年中可能错过的机器学习项目啦!
推荐阅读文章
在看排名清单之前,建议首先看看下面2篇文章:
A)神经网络
深度学习A-Z:动手的人工神经网络(Deep Learning A-Z: Hands-On Artificial Neural Networks)http://bit.ly/2CH1WcQ
[68,745条推荐,4.5 / 5星]
B)TensorFlow
用Python进行深度学习的TensorFlow完整指南(Complete Guide to TensorFlow for Deep Learning with Python)http://bit.ly/2EatVy7
[17,834条推荐,4.6 / 5星]
开源项目排行榜
第 1 名
FastText:用于快速文本表示和分类的库(FastText: Library for fast text representation and classification)。https://goo.gl/VWGfCs
[Github上的11786颗星] 由Facebook Research提供
...[Muse: 基于快速文本的多语言无监督或监督词嵌入(Muse: Multilingual Unsupervised or Supervised word Embeddings, based on Fast Text)。
Github上的695颗星] https://goo.gl/stbhjZ
第 2 名
深色照片风格转换:论文“深度照片风格转移”代码和数据(Deep-photo-styletransfer: Code and data for paper “Deep Photo Style Transfer”) https://github.com/luanfujun/deep-photo-styletransfer?utm_source=mybridge&utm_medium=blog&utm_campaign=read_more
[Github上的9747颗星星] 由康奈尔大学 Fujun Luan 博士提供
第 3 名
Python和世界上最简单的面部识别api和命令行(The world’s simplest facial recognition api for Python and the command line)https://github.com/ageitgey/face_recognition?utm_source=mybridge&utm_medium=blog&utm_campaign=read_more
[Github上8672个星] 由Adam Geitgey提供
第 4 名
洋红(Magenta: ):机器智能生成音乐和艺术(Magenta: Music and Art Generation with Machine Intelligence)。https://github.com/tensorflow/magenta?utm_source=mybridge&utm_medium=blog&utm_campaign=read_more
[GitHub上 8113颗星]
第 5 名
Sonnet:基于TensorFlow的神经网络库(Sonnet: TensorFlow-based neural network library)https://github.com/deepmind/sonnet?utm_source=mybridge&utm_medium=blog&utm_campaign=read_more
[Github上5731颗星]。 由Deepmind公司的Malcolm Reynolds(https://medium.com/@malcolm.reynolds)提供
第 6 名
deeplearn.js:为Web提供硬件加速的机器智能库(https://github.com/PAIR-code/deeplearnjs?utm_source=mybridge&utm_medium=blog&utm_campaign=read_more
[Github上5462颗星]。 由Google Brain的Nikhil Thorat提供
第 7 名
TensorFlow的快速风格迁移(Fast Style Transfer in TensorFlow)https://github.com/lengstrom/fast-style-transfer?utm_source=mybridge&utm_medium=blog&utm_campaign=read_more
[Github上4843颗星] 由MIT的Logan Engstrom提供
第 8 名
Pysc2:星际争霸II学习环境(Pysc2: StarCraft II Learning Environment)https://github.com/deepmind/pysc2?utm_source=mybridge&utm_medium=blog&utm_campaign=read_more
[Github上3683颗星星] 由DeepMind的Timo Ewalds提供
第 9 名
AirSim:基于Unreal Engine的开源模拟器,用于微软AI&Research自主车辆(AirSim: Open source simulator based on Unreal Engine for autonomous vehicles from Microsoft AI & Research)https://github.com/Microsoft/AirSim?utm_source=mybridge&utm_medium=blog&utm_campaign=read_more
[Github上3861颗星]。 由Microsoft的Shital Shah提供
第 10 名
Facets:机器学习数据集的可视化(Facets: Visualizations for machine learning datasets)https://github.com/PAIR-code/facets?utm_source=mybridge&utm_medium=blog&utm_campaign=read_more
[Github上3371颗星] 由Google Brain提供
第 11 名
Style2Paints:图像的AI彩色化(Style2Paints: AI colorization of images)https://github.com/lllyasviel/style2paints?utm_source=mybridge&utm_medium=blog&utm_campaign=read_more
[Github上3310颗星]
第 12 名
Tensor2Tensor:Google Research的广义序列到序列模型库 (Tensor2Tensor: A library for generalized sequence to sequence models — Google Research )https://github.com/tensorflow/tensor2tensor?utm_source=mybridge&utm_medium=blog&utm_campaign=read_more
[Github上3087颗星] 由Google Brain的Ryan Sepassi提供
第 13 名
PyTorch中的图像到图像转换(例如horse2zebra,edges2cats等)[Image-to-image translation in PyTorch (e.g. horse2zebra, edges2cats, and more)] https://github.com/junyanz/pytorch-CycleGAN-and-pix2pix?utm_source=mybridge&utm_medium=blog&utm_campaign=read_more
[Github上2847颗星] 由伯克利的Jun-Yan Zhu博士提供
第 14 名
Faiss:用于密集向量的高效相似性搜索和聚类的库(Faiss: A library for efficient similarity search and clustering of dense vectors)https://github.com/facebookresearch/faiss?utm_source=mybridge&utm_medium=blog&utm_campaign=read_more
[Github上2629颗星]。 由Facebook Research提供
第 15 名
Fashion-mnist:类似MNIST的时尚产品数据库(Fashion-mnist: A MNIST-like fashion product database)https://github.com/zalandoresearch/fashion-mnist?utm_source=mybridge&utm_medium=blog&utm_campaign=read_more
[Github上2780颗星星] 由Zalando Tech的研究科学家Han Xiao提供
第 16 名
ParlAI:在各种公开可用的对话数据集上训练和评估AI模型的框架(https://github.com/facebookresearch/ParlAI?utm_source=mybridge&utm_medium=blog&utm_campaign=read_more)
[Github上2578颗星]。 有Facebook研究院的Alexander Miller提供
第 17 名
Fairseq:Facebook AI研究序列 - 序列工具包(Fairseq: Facebook AI Research Sequence-to-Sequence Toolkit)https://github.com/facebookresearch/fairseq?utm_source=mybridge&utm_medium=blog&utm_campaign=read_more
[Github上2571颗星]
第 18 名
Pyro:Python和PyTorch的深度通用概率编程(Pyro: Deep universal probabilistic programming with Python and PyTorch)https://github.com/uber/pyro?utm_source=mybridge&utm_medium=blog&utm_campaign=read_more
[Github上的2387颗星] 由Uber Engineering提供
第 19 名
iGAN:由GAN支持的交互式图像生成(iGAN: Interactive Image Generation powered by GAN)https://github.com/junyanz/iGAN?utm_source=mybridge&utm_medium=blog&utm_campaign=read_more
[Github上2369颗星]
第 20 名
Deep-image-prior:使用无需学习的神经网络进行图像恢复(Deep-image-prior: Image restoration with neural networks but without learning)https://github.com/DmitryUlyanov/deep-image-prior?utm_source=mybridge&utm_medium=blog&utm_campaign=read_more
[Github上2188颗星] 由Skoltech博士Dmitry Ulyanov提供
第 21 名
Face_classification:使用具有keras CNN模型和openCV的fer2013 / imdb数据集的实时人脸检测和情感/性别分类(Face_classification: Real-time face detection and emotion/gender classification using fer2013/imdb datasets with a keras CNN model and openCV)https://github.com/oarriaga/face_classification?utm_source=mybridge&utm_medium=blog&utm_campaign=read_more
[Github上1967星]
第 22 名
Speech-to-Text-WaveNet:使用DeepMind的WaveNet和tensorflow进行端到端的句子级英语语音识别(Speech-to-Text-WaveNet : End-to-end sentence level English speech recognition using DeepMind’s WaveNet and tensorflow )https://github.com/buriburisuri/speech-to-text-wavenet?utm_source=mybridge&utm_medium=blog&utm_campaign=read_more
[Github上1961星] 由Kakao Brain的Namju Kim提供
第 23 名
StarGAN:用于多域图像到图像转换的统一生成对抗网络(StarGAN: Unified Generative Adversarial Networks for Multi-Domain Image-to-Image Translation)https://github.com/yunjey/StarGAN?utm_source=mybridge&utm_medium=blog&utm_campaign=read_more
[Github上1954颗星] 由韩国大学的Yunjey Choi提供
第 24 名
Ml-agents:Unity机器学习代理(Ml-agents: Unity Machine Learning Agents)
[Github上的1658颗星] Deep Learning at Unity3D 的Unity Juliani提供
第 25 名
DeepVideoAnalytics:分布式可视化搜索和可视化数据分析平台(DeepVideoAnalytics: A distributed visual search and visual data analytics platform)https://github.com/AKSHAYUBHAT/DeepVideoAnalytics/?utm_source=mybridge&utm_medium=blog&utm_campaign=read_more
[Github上1494颗星] 由康奈尔大学的Akshay Bhat博士提供
第 26 名
OpenNMT:火炬中的开源神经机器翻译(OpenNMT: Open-Source Neural Machine Translation in Torch)https://github.com/OpenNMT/OpenNMT?utm_source=mybridge&utm_medium=blog&utm_campaign=read_more
[Github上1490颗星]
第 27 名
Pix2pixHD:使用条件GAN合成和处理2048×1024的图像(Pix2pixHD: Synthesizing and manipulating 2048x1024 images with conditional GANs )https://github.com/NVIDIA/pix2pixHD?utm_source=mybridge&utm_medium=blog&utm_campaign=read_more
[Github上1283颗星] 由Nvidia的AI Research 科学家 Ming-Yu Liu提供
第 28 名
Horovod:TensorFlow的分布式训练框架(Horovod: Distributed training framework for TensorFlow) https://github.com/uber/horovod?utm_source=mybridge&utm_medium=blog&utm_campaign=read_more
[Github上1188颗星] 由Uber Engineering提供
第 29 名
AI-Blocks:强大的和直观的所见即所得的界面,允许任何人创建的机器学习模型(AI-Blocks: A powerful and intuitive WYSIWYG interface that allows anyone to create Machine Learning models)https://github.com/MrNothing/AI-Blocks?utm_source=mybridge&utm_medium=blog&utm_campaign=read_more
[GitHub上899颗星]
第 30 名
采用Tensorflow的深度神经网络进行语音转换(语音风格迁移)[Deep neural networks for voice conversion (voice style transfer) in Tensorflow] https://github.com/andabi/deep-voice-conversion?utm_source=mybridge&utm_medium=blog&utm_campaign=read_more
[Github上845颗星] 有AI Kakao Brain的AI 研究院的Dabi Ahn提供
参考文章:
30 Amazing Machine Learning Projects for the Past Year
https://medium.mybridge.co/30-amazing-machine-learning-projects-for-the-past-year-v-2018-b853b8621ac7
领取专属 10元无门槛券
私享最新 技术干货