Mybridge发布了一篇文章,对比了过去一年中机器学习领域大约8800个开源项目后,选出30个2017年度优秀的开源项目,包含机器学习开源库、数据库以及其他应用程序,这些项目差不多都是在2017年1-12月发布。MybridgeAI通过受欢迎度、参与度以及其他方面对开源项目进行评定。
对于机器学习者来说,阅读开源代码并基于代码构建自己的项目,是一个非常有效的学习方法。看看以下这些Github上平均star为3558的开源项目,你错了哪些?
在开始之前,先推荐阅读:
A. 神经网络:深度学习 A-ZTM : 亲手搭建人工神经网络(推荐次数68,745 , 4.5/5 stars)
B. 用Python进行深度学习的TensorFlow的完整指南(推荐次数17,834, 4.6/5 stars)
接下来是Mybridge精选的Top30的项目:
源码链接:https://github.com/facebookresearch/MUSE
源码链接:https://github.com/luanfujun/deep-photo-styletransfer
源码链接:https://github.com/ageitgey/face_recognition
源码链接:https://github.com/tensorflow/magenta
源码链接:https://github.com/deepmind/sonnet
源码链接:https://github.com/PAIR-code/deeplearnjs
源码链接:https://github.com/lengstrom/fast-style-transfer
源码链接:https://github.com/deepmind/pysc2
源码链接:https://github.com/Microsoft/AirSim
源码链接:https://github.com/PAIR-code/facets
源码链接:https://github.com/lllyasviel/style2paints
源码链接:https://github.com/tensorflow/tensor2tensor
源码地址:https://github.com/junyanz/pytorch-CycleGAN-and-pix2pix
源码地址:https://github.com/facebookresearch/faiss
源码链接:https://github.com/zalandoresearch/fashion-mnist
源码链接:https://github.com/facebookresearch/ParlAI
源码链接:https://github.com/facebookresearch/fairseq
源码链接:https://github.com/uber/pyro
源码地址:https://github.com/junyanz/iGAN
源码地址:https://github.com/DmitryUlyanov/deep-image-prior
源码地址:https://github.com/oarriaga/face_classification
源码地址:https://github.com/buriburisuri/speech-to-text-wavenet
源码地址:https://github.com/yunjey/StarGAN
源码地址:https://github.com/Unity-Technologies/ml-agents
源码地址:https://github.com/AKSHAYUBHAT/DeepVideoAnalytics
源码地址:https://github.com/OpenNMT/OpenNMT
源码地址:https://github.com/NVIDIA/pix2pixHD
源码地址:https://github.com/uber/horovod
源码地址:https://github.com/MrNothing/AI-Blocks