深度学习入门教程
新手教程内含四个板块,分别为网络课程,常用文档,技术博客,常用网上资源。四个板块中的链接都是主流的深度学习中强有力的学习资料。
目录
UFLDL和cs231n并行的看,一边做练习一边看理论。优先实践,多看论文,多写代码,可以先用keras上手, 之后转tensorflow/pytorch
1. 网络课程
- UFLDL: http://ufldl.stanford.edu/tutorial/
- Stanford cs231n 官网链接: http://cs231n.stanford.edu/syllabus.html
- 一些论文: https://github.com/terryum/awesome-deep-learning-papers 其中Convolutional Neural Network Models这一节下的论文都要看,从旧往新看,看完这个看object detection的论文,然后看Seq2seq相关论文。 论文看到知道怎么实现才可以称为看懂,最好是自己实现一遍
tf tutorials: https://github.com/vahidk/EffectiveTensorflow
2.常用文档
Keras: https://keras.io https://github.com/fchollet/keras/tree/master/examples TensorFlow: https://www.tensorflow.org/
http://docs.w3cub.com/tensorflow~guide/ http://docs.w3cub.com/tensorflow~python/
3.技术博客
垠神的人生博客:http://www.yinwang.org/blog-cn/2017/04/23/ai
handong的图像博客:https://handong1587.github.io/deep_learning/2015/10/09/ocr.html
hetong的文字识别博客:http://tonghe90.github.io
何恺明博士的主页: http://kaiminghe.com/ (比较著名的作品有Faster-RCNN, Mask-RCNN, 残差网络, 暗通道先验去雾算法)
Yarin Gal Bayesian Neural Network: http://www.cs.ox.ac.uk/people/yarin.gal/website/ (异端统计学做得比较好的一个哥们)
Kevin Zakka 讲 Spatial Transformer Network: https://kevinzakka.github.io/2017/01/18/stn-part2/
讲tf入门与基础的blog: https://jacobbuckman.com/post/tensorflow-the-confusing-parts-1/
CV组的老朋友迪豪大神的知乎:https://www.zhihu.com/people/tobegit3hub/posts
如何退出vim:https://medium.freecodecamp.org/one-out-of-every-20-000-stack-overflow-visitors-is-just-trying-to-exit-vim-5a6b6175e7b6
巴黎最强CV实验室WILLOW:https://www.di.ens.fr/willow/research.php
深度学习500问:https://github.com/scutan90/DeepLearning-500-questions
SOAT 算法梳理: https://github.com/BlinkDL/BlinkDL.github.io
装机瞎倒腾:http://timdettmers.com/2018/12/16/deep-learning-hardware-guide/
4.常用网上资源
https://github.com/floodsung/Deep-Learning-Papers-Reading-Roadmap https://github.com/ChristosChristofidis/awesome-deep-learning#researchers https://github.com/kjw0612/awesome-deep-vision https://github.com/amusi/awesome-object-detection
gluon-cv: https://gluon-cv.mxnet.io/ https://github.com/dmlc/gluon-cv https://github.com/mrgloom/awesome-semantic-segmentation