1- # DeepLearning Tutorial
1+ # DeepLearning Tutorial
22## 一. 入门资料
33
4+ [ ** 完备的 AI 学习路线,最详细的中英文资源整理** ] ( https://zhuanlan.zhihu.com/p/64080846 )
5+
6+ ### 数学基础
7+
8+ ![ ] ( notes/Images/机器学习和深度学习的数学基础.png )
9+
410### 机器学习基础
511
612#### 快速入门
5662### 工程能力
5763* [ 机器学习算法实战] ( #机器学习实战篇 )
5864* [ 深度学习框架] ( #深度学习框架 )
65+ * [ 《AI算法工程师手册》] ( http://www.huaxiaozhuan.com/ )
5966* [ 算法工程师面试] ( https://github.com/imhuay/Algorithm_Interview_Notes-Chinese )
6067* [ 深度学习500问] ( https://github.com/scutan90/DeepLearning-500-questions )
61- * [ Kaggle实战] ( ) :[ 分分钟带你杀入Kaggle Top 1%] ( https://zhuanlan.zhihu.com/p/27424282 ) && [ 如何达到Kaggle竞赛top 2%?这里有一篇特征探索经验帖] ( https://zhuanlan.zhihu.com/p/48758045 ) && [ 如何在 Kaggle 首战中进入前 10%?] ( https://zhuanlan.zhihu.com/p/27486736 )
68+ * [ Kaggle实战] ( )
69+ * 常用算法:
70+ * Feature Engineering:continue variable && categorical variable
71+ * Classic machine learning algorithm:LR, KNN, SVM, Random Forest, GBDT(XGBoost&&LightGBM), Factorization Machine, Field-aware Factorization Machine, Neural Network
72+ * Cross validation, model selection:grid search, random search, hyper-opt
73+ * Ensemble learning
74+ * [ Kaggle入门系列:(一)机器学习环境搭建] ( https://zhuanlan.zhihu.com/p/29086448 ) && [ Kaggle入门系列:(二)Kaggle简介] ( https://zhuanlan.zhihu.com/p/29417603 ) && [ Kaggle入门系列(三)Titanic初试身手] ( https://zhuanlan.zhihu.com/p/29086614 )
75+ * [ 从 0 到 1 走进 Kaggle] ( https://zhuanlan.zhihu.com/p/61660061 )
76+ * [ Kaggle 入门指南] ( https://zhuanlan.zhihu.com/p/25742261 )
77+ * [ 一个框架解决几乎所有机器学习问题] ( https://zhuanlan.zhihu.com/p/61657532 ) && [ Approaching (Almost) Any Machine Learning Problem | Abhishek Thakur] ( http://blog.kaggle.com/2016/07/21/approaching-almost-any-machine-learning-problem-abhishek-thakur/ )
78+ * [ 分分钟带你杀入Kaggle Top 1%] ( https://zhuanlan.zhihu.com/p/27424282 )
79+ * [ 如何达到Kaggle竞赛top 2%?这里有一篇特征探索经验帖] ( https://zhuanlan.zhihu.com/p/48758045 )
80+ * [ 如何在 Kaggle 首战中进入前 10%?] ( https://zhuanlan.zhihu.com/p/27486736 )
6281
6382## 二. 神经网络模型概览
6483
100119 * [ 人脸识别算法演化史] ( https://zhuanlan.zhihu.com/p/36416906 )
101120 * [ CascadeCNN] ( https://blog.csdn.net/shuzfan/article/details/50358809 )
102121 * [ MTCNN] ( https://blog.csdn.net/qq_14845119/article/details/52680940 )
122+ * [ awesome-Face_Recognition] ( https://github.com/ChanChiChoi/awesome-Face_Recognition )
123+ * [ 异质人脸识别研究综述] ( https://zhuanlan.zhihu.com/p/64191484 )
124+ * [ 老板来了:人脸识别+手机推送,老板来了你立刻知道。] ( https://zhuanlan.zhihu.com/p/26431250 ) && [ 手把手教你用Python实现人脸识别] ( https://zhuanlan.zhihu.com/p/33456076 ) && [ 人脸识别项目,网络模型,损失函数,数据集相关总结] ( https://www.jianshu.com/p/e57205edc364 )
125+ * [ 基于深度学习的人脸识别技术综述] ( https://zhuanlan.zhihu.com/p/24816781 ) && [ 如何走近深度学习人脸识别?你需要这篇超长综述] ( https://zhuanlan.zhihu.com/p/35295839 ) && [ 人脸识别损失函数综述(附开源实现)] ( https://zhuanlan.zhihu.com/p/51324547 )
103126* [ 深度学习图像超分辨率综述] ( https://zhuanlan.zhihu.com/p/57564211 )
104127* [ 目标检测进化史] ( https://zhuanlan.zhihu.com/p/60590369 )
105128* [ 一文看尽21篇目标检测最新论文(腾讯/Google/商汤/旷视/清华/浙大/CMU/华科/中科院等] ( https://zhuanlan.zhihu.com/p/61080508 )
106129* [ Anchor-Free目标检测算法] ( ) : [ 第一篇:arxiv2015_baidu_DenseBox] ( https://zhuanlan.zhihu.com/p/40221183 ) , [ 如何评价最新的anchor-free目标检测模型FoveaBox?] ( https://www.zhihu.com/question/319605567/answer/647844997 ) , [ FCOS: 最新的one-stage逐像素目标检测算法] ( https://zhuanlan.zhihu.com/p/61644900 ) && [ 最新的Anchor-Free目标检测模型FCOS,现已开源!] ( https://zhuanlan.zhihu.com/p/62198865 ) && [ 中科院牛津华为诺亚提出CenterNet,one-stage detector可达47AP,已开源!] ( https://zhuanlan.zhihu.com/p/62789701 )
107130* [ 目标检测算法综述之FPN优化篇] ( https://zhuanlan.zhihu.com/p/62975854 )
108131* [ https://zhuanlan.zhihu.com/p/63273342 ] ( 聊聊Anchor的"前世今生"(上) )
132+ <<<<<<< HEAD
109133* [ 【CVPR2019正式公布】行人重识别论文] ( https://zhuanlan.zhihu.com/p/62843442 ) ,[ 2019 行人再识别年度进展回顾] ( https://zhuanlan.zhihu.com/p/64004977 )
110134* [ 从SRCNN到EDSR,总结深度学习端到端超分辨率方法发展历程] ( https://zhuanlan.zhihu.com/p/31664818 )
135+ =======
136+ * [ 【CVPR2019正式公布】行人重识别论文] ( https://zhuanlan.zhihu.com/p/62843442 )
137+ * [ 自然场景文本检测识别技术综述] ( https://mp.weixin.qq.com/s?__biz=MzU4MjQ3MDkwNA==&mid=2247485142&idx=1&sn=c0e01da30eb5e750be453eabe4be2bf4&chksm=fdb69b41cac11257ae22c7dac395e9651dab628fc35dd6d3c02d9566a8c7f5f2b56353d58a64&token=1065243837&lang=zh_CN#rd )
138+ >>>>>>> 4942f011e134a9b9b77d78755ed8f06ddcf91462
111139
112140#### 教程
113141
169197
170198### GAN
171199
200+ * [ 苏剑林博客,讲解得淋漓尽致] ( https://kexue.fm/category/Big-Data )
201+
172202#### 发展史
173203
174204* [ 千奇百怪的GAN变体] ( https://zhuanlan.zhihu.com/p/26491601 )
267297* [ 20. BFGS] ( https://blog.csdn.net/philosophyatmath/article/details/70173128 )
268298* [ 21. 详解深度学习中的梯度消失、爆炸原因及其解决方法] ( https://zhuanlan.zhihu.com/p/33006526 )
269299* [ 22. Dropout] ( https://arxiv.org/pdf/1207.0580.pdf ) , [ 1] ( https://blog.csdn.net/stdcoutzyx/article/details/49022443 ) , [ 2] ( https://blog.csdn.net/hjimce/article/details/50413257 ) , [ 3] ( https://blog.csdn.net/shuzfan/article/details/50580915 )
300+ * [ 23.谱归一化(Spectral Normalization)的理解] ( https://blog.csdn.net/StreamRock/article/details/83590347 ) ,[ 常见向量范数和矩阵范数] ( https://blog.csdn.net/left_la/article/details/9159949 ) ,[ 谱范数正则(Spectral Norm Regularization)的理解] ( https://blog.csdn.net/StreamRock/article/details/83539937 )
301+ * [ 24.L1正则化与L2正则化] ( https://zhuanlan.zhihu.com/p/35356992 )
302+ * [ 25.为什么选用交叉熵而不是MSE] ( https://zhuanlan.zhihu.com/p/61944055 )
270303
271304## 四. 炼丹术士那些事
272305
275308* [ 神经网络训练trick] ( https://zhuanlan.zhihu.com/p/59918821 )
276309* [ 深度学习与计算机视觉系列(8)_ 神经网络训练与注意点] ( https://blog.csdn.net/han_xiaoyang/article/details/50521064 )
277310* [ 神经网络训练loss不下降原因集合] ( https://blog.csdn.net/liuweiyuxiang/article/details/80856991 )
278- * [ 深度学习:欠拟合问题的几种解决方案] ( https://blog.csdn.net/ningyanggege/article/details/82183666 )
311+ * [ 深度学习:欠拟合问题的几种解决方案] ( https://blog.csdn.net/ningyanggege/article/details/82183666 ) && [ 过拟合和欠拟合问题 ] ( https://blog.csdn.net/mzpmzk/article/details/79741682 )
279312* [ 机器学习:如何找到最优学习率] ( https://blog.csdn.net/whut_ldz/article/details/78882871 ) 及[ 实现] ( https://github.com/L1aoXingyu/torchlib )
280313* [ 不平衡数据集处理方法] ( ) : [ 其一] ( https://machinelearningmastery.com/tactics-to-combat-imbalanced-classes-in-your-machine-learning-dataset/ ) , [ 其二] ( https://www.zhihu.com/question/285824343 ) , [ 其三] ( https://blog.csdn.net/songhk0209/article/details/71484469 )
281314* [ 同一个神经网络使用不同激活函数的表达能力是否一致] ( https://www.zhihu.com/question/41841299 )
340373
341374### 深度学习框架
342375
343- #### Python3.x(前提)
376+ #### Python3.x(先修)
377+
378+ * [ The Python Tutorial] ( https://docs.python.org/3/tutorial/ )
379+
344380* [ 廖雪峰Python教程] ( https://www.liaoxuefeng.com/wiki/0014316089557264a6b348958f449949df42a6d3a2e542c000 )
345381* [ 菜鸟教程] ( http://www.runoob.com/python3/python3-tutorial.html )
382+ * [ 给深度学习入门者的Python快速教程 - 基础篇] ( https://zhuanlan.zhihu.com/p/24162430 )
383+
384+ #### Numpy(先修)
385+
386+ * [ Quickstart tutorial] ( https://www.numpy.org/devdocs/user/quickstart.html )
387+
388+ * [ Numpy快速入门(Numpy 1.14 官方文档中文翻译)] ( https://www.jianshu.com/p/3e566f09a0cf )
389+ * [ Numpy中文文档] ( https://www.numpy.org.cn/index.html )
390+ * [ 给深度学习入门者的Python快速教程 - numpy和Matplotlib篇] ( https://zhuanlan.zhihu.com/p/24309547 )
391+
392+ #### Opencv-python
393+
394+ * [ OpenCV-Python Tutorials] ( https://opencv-python-tutroals.readthedocs.io/en/latest/py_tutorials/py_tutorials.html )
395+ * [ OpenCV官方教程中文版(For Python)] ( https://www.cnblogs.com/Undo-self-blog/p/8423851.html )
396+ * [ 数字图像处理系列] ( https://blog.csdn.net/feilong_csdn/article/category/8037591 )
397+ * [ python+OpenCV图像处理] ( https://blog.csdn.net/qq_40962368/article/category/7688903 )
398+ * [ 给深度学习入门者的Python快速教程 - 番外篇之Python-OpenCV] ( https://zhuanlan.zhihu.com/p/24425116 )
399+
400+ #### Pandas
401+
402+ * [ Python 数据科学入门教程:Pandas] ( https://www.jianshu.com/p/d9774cf1fea5?utm_campaign=maleskine&utm_content=note&utm_medium=seo_notes&utm_source=recommendation )
346403
347404#### Tensorflow
348405
349406* [ 如何高效地学习 TensorFlow 代码] ( https://www.zhihu.com/question/41667903 )
350-
351407* [ 中文教程] ( http://www.tensorfly.cn/tfdoc/tutorials/overview.html )
352408* [ TensorFlow官方文档] ( https://www.w3cschool.cn/tensorflow_python/ )
353409* [ CS20: Tensorflow for DeepLearning Research] ( http://web.stanford.edu/class/cs20si/syllabus.html )
354410* [ 吴恩达TensorFlow专项课程] ( https://zhuanlan.zhihu.com/p/62981537 )
355411* [ 【干货】史上最全的Tensorflow学习资源汇总] ( https://zhuanlan.zhihu.com/p/35515805?group_id=967136289941897216 )
356412* [ 《21个项目玩转深度学习———基于TensorFlow的实践详解》] ( https://github.com/hzy46/Deep-Learning-21-Examples )
413+ * [ 最全Tensorflow2.0 入门教程持续更新] ( https://zhuanlan.zhihu.com/p/59507137 )
357414
358415#### MXNet
359416* [ Gluon] ( http://zh.gluon.ai/# )
363420#### PyTorch
364421* [ PyTorch中文文档] ( https://pytorch-cn.readthedocs.io/zh/latest/ )
365422* [ WELCOME TO PYTORCH TUTORIALS] ( https://pytorch.org/tutorials/index.html )
423+ * [ 史上最全的PyTorch学习资源汇总] ( https://zhuanlan.zhihu.com/p/64895011 )
366424
367425### Python可视化
368426* [ Top 50 matplotlib Visualizations – The Master Plots (with full python code)] ( https://www.machinelearningplus.com/plots/top-50-matplotlib-visualizations-the-master-plots-python/ )
427+ * [ Python之MatPlotLib使用教程] ( https://www.jianshu.com/p/92e1a4497505 )
428+ * [ 给深度学习入门者的Python快速教程 - numpy和Matplotlib篇] ( https://zhuanlan.zhihu.com/p/24309547 )
369429
370430### 标注工具
371431* 目标检测标注工具
@@ -494,6 +554,9 @@ ______
494554
495555* [ 推荐算法相关的文档整理] ( https://zhuanlan.zhihu.com/p/29969721 )
496556* [ Embedding从入门到专家必读的十篇论文] ( https://zhuanlan.zhihu.com/p/58805184 )
557+ * 推荐系统之路
558+ * [ 推荐系统之路 (1):走上推荐系统这条路] ( https://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650760136&idx=2&sn=afc75d6bf614bc7929b6ea9cb1abb260&chksm=871aa7b6b06d2ea0129ec7b06bf7b2448c3a55d485d6b80a066d622709066242fe7c925160c3&scene=21#wechat_redirect )
559+ * [ 推荐系统之路 (2):产品聚类] ( https://zhuanlan.zhihu.com/p/64722876 )
497560
498561### 自然语言处理(NLP)
499562
@@ -522,7 +585,12 @@ ______
522585
523586### 机器学习理论篇
524587
588+ #### 逻辑回归
589+
590+ * [ 【机器学习面试题】逻辑回归篇] ( https://zhuanlan.zhihu.com/p/62653034 )
591+
525592#### 决策树(Decision Tree)
593+
526594* [ Python3《机器学习实战》学习笔记(二):决策树基础篇之让我们从相亲说起] ( https://blog.csdn.net/c406495762/article/details/75663451 )
527595* [ Python3《机器学习实战》学习笔记(三):决策树实战篇之为自己配个隐形眼镜] ( https://blog.csdn.net/c406495762/article/details/76262487 )
528596* [ 机器学习实战教程(十三):树回归基础篇之CART算法与树剪枝] ( http://cuijiahua.com/blog/2017/12/ml_13_regtree_1.html )
@@ -532,6 +600,10 @@ ______
532600* [ 决策树值ID3、C4.5实现] ( https://blog.csdn.net/u014688145/article/details/53212112 )
533601* [ 决策树值CART实现] ( https://blog.csdn.net/u014688145/article/details/53326910 )
534602
603+ #### 随机森林
604+
605+ * [ 随机森林(Random Forest)入门与实战] ( https://blog.csdn.net/sb19931201/article/details/52601058 )
606+
535607#### 支持向量机(SVM)
536608* [ SVM通俗导论 July(本文章是我看过最好的SVM导论)] ( https://github.com/Mikoto10032/DeepLearning/blob/master/books/%E6%94%AF%E6%8C%81%E5%90%91%E9%87%8F%E6%9C%BA%E9%80%9A%E4%BF%97%E5%AF%BC%E8%AE%BA%EF%BC%88%E7%90%86%E8%A7%A3SVM%E7%9A%84%E4%B8%89%E5%B1%82%E5%A2%83%E7%95%8C%EF%BC%89LaTeX%E6%9C%80%E6%96%B0%E7%89%88_2015.1.9.pdf )
537609* [ Python3《机器学习实战》学习笔记(八):支持向量机原理篇之手撕线性SVM (SMO训练过程总结得清晰易懂)] ( https://blog.csdn.net/c406495762/article/details/78072313 )
@@ -574,10 +646,19 @@ ______
574646* [ 梯度提升决策树] ( https://zhuanlan.zhihu.com/p/36339161 )
575647* [ GBDT原理及应用] ( https://zhuanlan.zhihu.com/p/30339807 )
576648* [ XGBOOST原理篇] ( https://zhuanlan.zhihu.com/p/31654000 )
649+ * [ xgboost入门与实战(原理篇)] ( https://blog.csdn.net/sb19931201/article/details/52557382 ) && [ xgboost入门与实战(实战调参篇)] ( https://blog.csdn.net/sb19931201/article/details/52577592 )
650+ * [ 【干货合集】通俗理解kaggle比赛大杀器xgboost] ( https://zhuanlan.zhihu.com/p/41417638 )
651+ * [ GBDT分类的原理及Python实现] ( https://blog.csdn.net/bf02jgtrs00xktcx/article/details/82719765 )
652+ * [ GBDT原理及利用GBDT构造新的特征-Python实现] ( https://blog.csdn.net/shine19930820/article/details/71713680 )
653+ * [ Python+GBDT算法实战——预测实现100%准确率] ( https://www.jianshu.com/p/47e73a985ba1 )
654+
655+ #### 集成(Ensemble)
656+
657+ - [ 集成学习法之bagging方法和boosting方法] ( https://blog.csdn.net/qq_30189255/article/details/51532442 )
658+ - [ Bagging,Boosting,Stacking] ( https://blog.csdn.net/Mr_tyting/article/details/72957853 )
659+
660+ #### EM(期望最大化)
577661
578- #### 集成(EM)
579- * [ 集成学习法之bagging方法和boosting方法] ( https://blog.csdn.net/qq_30189255/article/details/51532442 )
580- * [ Bagging,Boosting,Stacking] ( https://blog.csdn.net/Mr_tyting/article/details/72957853 )
581662* [ 人人都懂的EM算法 ] ( https://zhuanlan.zhihu.com/p/36331115 )
582663* [ EM算法入门文章] ( https://zhuanlan.zhihu.com/p/61768577 )
583664
0 commit comments