This repository has been archived by the owner on Jul 10, 2023. It is now read-only.
-
Notifications
You must be signed in to change notification settings - Fork 0
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
- Loading branch information
Showing
12 changed files
with
208 additions
and
18 deletions.
There are no files selected for viewing
This file was deleted.
Oops, something went wrong.
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,3 @@ | ||
## TODOS | ||
|
||
- [ ] 研究优秀的深度学习资源 |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,22 @@ | ||
### 个人 Python 项目 | ||
|
||
1 自制数据集 X:https://github.com/DataLoaderX/datazone | ||
|
||
- Situation: 创建自定义的数据集以适用于不同的深度学习框架。 | ||
- Task: 设计一个 API 将 MNIST, Cifar 10, cifar100, Fashion-MNIST 封装为数据集 X,并将 X 其保存为 HDF5 格式。 | ||
- Action: 借助 Python 的 Bunch 结构将数据进行管理,将图片以 Numpy 的形式进行封装,之后可以直接以数组的形式获取图片。将各个数据集的标签名称、数值标签以及数据集的源网站都封装进入数据集 X。 | ||
- Result: HDF5 是一个可以高效的存储和读取的数据结构,不仅仅支持 Python,也支持 Matlab 获取(https://yq.aliyun.com/articles/614332?spm=a2c4e.11155435.0.0.30543312vFsboY)。数据 X 被我被我放在了多个平台(博客园,云栖社区,简书)进行分享都获得超过 400 的阅读量。 | ||
|
||
2 改写 cocoapi:https://github.com/Xinering/cocoapi | ||
|
||
- Situation: 利用 Python 的特性对原微软的 api 进行改写使其支持直接读取压缩文件而直接跳过繁琐的解压和重写工作。 | ||
- Task:利用 Python 的类的继承机制改写 cocoapi,并命名为 cocoz,令 cocoz 可以更加方便的处理 COCO 数据集,并且还可以使用 cocoz 来处理类似于 COCO 数据集的形式的数据集(https://nbviewer.jupyter.org/github/XinetAI/CVX/blob/master/Notebook/dog_cat_gluon.ipynb)。 | ||
- Action:创建 cocoz.py 来完成预期任务,并创建 pycocoZDemo.ipynb 和 pycocoZEvalDemo.ipynb 来说明如何使用 cocoz.py。 | ||
- Result:将 cocoz 的创建与使用分享在博客园(https://www.cnblogs.com/q735613050/p/8969452.html) 已经获得 15816 的阅读量(2019/3/22)。与北京源智天下科技有限公司签约写关于如何动手实现计算机视觉的书。 | ||
|
||
3 创建处理脱机和在线手写汉字库的 API:https://github.com/DataLoaderX/datasetsome | ||
|
||
- Situation:中科院自动化研究所在 2007-2010 年间收集的 CASIA-HWDB 和 CASIA-OLHWDB 数据集。该数据集的学术研究的用途包括:手写文档分割、字符识别、字符串识别、文档检索、书写人适应、书写人鉴别等。但是直接使用该数据库有点困难,需要详细阅读官方提供的文档说明,并且辅以各种编码知识。 | ||
- Task:简化手写汉字库的读取过程,令数据的载入和读取更加人性化。 | ||
- Action:创建 xhw.py 实现数据的封装,以 HDF 格式直接获取手写汉字的图片和特征信息。 | ||
- Result:在慕课网分享该 API 获取超过 1000 的阅读量。 |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,44 @@ | ||
### 1 优质知乎专栏 | ||
|
||
- [深度炼丹](https://zhuanlan.zhihu.com/c_94953554) | ||
- [From Beijing with Love](https://zhuanlan.zhihu.com/frombeijingwithlove) | ||
- [使用MXNet/Gluon来动手学深度学习](https://zhuanlan.zhihu.com/gluon) | ||
- [深度炼丹](https://zhuanlan.zhihu.com/c_94953554) | ||
|
||
### 2 优质教程 | ||
|
||
- [中文版 Gluon 教程](https://zh.diveintodeeplearning.org/) | ||
|
||
### 3 学习 | ||
|
||
- [学习材料](学习材料.md) | ||
- [Linux工具快速教程](https://linuxtools-rst.readthedocs.io/zh_CN/latest/) | ||
- [Awesome TensorFlow](https://github.com/jtoy/awesome-tensorflow) | ||
- [实战 Kaggle 比赛:房价预测](https://zh.diveintodeeplearning.org/chapter_deep-learning-basics/kaggle-house-price.html) | ||
- [Face Recognition Project on MXNet](https://github.com/deepinsight/insightface) | ||
|
||
论文阅读路线图【hard way】: github上有个很有名的 [Deep Learning papers reading roadmap](https://github.com/floodsung/Deep-Learning-Papers-Reading-Roadmap),已经过万stars,从事深度学习科研者的必读清单。 | ||
|
||
### 4 框架 | ||
|
||
- [MXNet](https://mxnet.apache.org/) | ||
- [Rapid Enterprise Mobile Apps Build, Connect, Manage.](https://gluonhq.com/) | ||
|
||
### 5 优质博客 | ||
|
||
1. [Beginner-level tutorials for a powerful framework](http://learningtensorflow.com/) | ||
2. [CS231n: Convolutional Neural Networks for Visual Recognition](http://cs231n.stanford.edu/) | ||
3. [CS224n: Natural Language Processing with Deep Learning](http://web.stanford.edu/class/cs224n/) | ||
4. [UFLDL Tutorial](http://deeplearning.stanford.edu/tutorial/) | ||
5. [UFLDL教程(上面UFLDL Tutorial旧版,但是是中文的)](http://deeplearning.stanford.edu/wiki/index.php/UFLDL%E6%95%99%E7%A8%8B) | ||
6. [First Contact With TensorFlow](http://jorditorres.org/first-contact-with-tensorflow/) | ||
7. [Deep Learning](http://deeplearning.net/) | ||
8. [colah’s blog(推荐,写了很多晦涩概念可视化的博客)](http://colah.github.io/) | ||
9. [DeepLearning4J(虽然是一个面向Java的深度学习框架,但是文档很好,通用性也很强)](https://deeplearning4j.org/) | ||
10. [【干货合集】Docker快速入门与进阶](https://cloud.tencent.com/developer/article/1004996) | ||
|
||
### 6 不错的资源类网站 | ||
|
||
1. [arxiv-sanity](http://www.arxiv-sanity.com/) | ||
2. [GitXiv](http://www.gitxiv.com/) | ||
3. [mybinder](https://mybinder.org/) 将 GitHub 上的 Jupyter NoteBook 推送至云端 |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,105 @@ | ||
### 不断筛选优秀资料: | ||
|
||
- [MXNet 中文教程:图像分类](http://blog.csdn.net/ironyoung/article/details/53968176) | ||
- [用MXnet实战深度学习之一:安装GPU版mxnet并跑一个MNIST手写数字识别](http://www.open-open.com/lib/view/open1448030000650.html) | ||
- [dmlc/mxnet-notebooks](https://github.com/dmlc/mxnet-notebooks) | ||
- [MXNet API入门 —第1篇](http://www.infoq.com/cn/articles/an-introduction-to-the-mxnet-api-part01) | ||
- Apache MXNet官方网站:https://mxnet.incubator.apache.org/ | ||
- 安装指南:https://mxnet.incubator.apache.org/versions/master/get_started/install.html | ||
- Gluon教程:http://gluon.mxnet.io/ | ||
- Gluon讲座PPT: https://github.com/zackchase/mxnet-slides/ | ||
- Gluon深度学习样例:https://github.com/apache/incubator-mxnet/tree/master/example/gluon | ||
- Gluon 论坛:https://discuss.gluon.ai/ | ||
- Gluon 中文教程:http://zh.gluon.ai/index.html | ||
- [gluon:使用MXNet/Gluon来动手学深度学习](https://zhuanlan.zhihu.com/gluon) | ||
- [YOLO,一种简易快捷的目标检测算法](https://zhuanlan.zhihu.com/p/32945351) | ||
- [从零开始码一个皮卡丘检测器-CNN目标检测入门教程](https://zhuanlan.zhihu.com/p/28867241) | ||
- [用Gluon炼丹体验](https://zhuanlan.zhihu.com/p/30966663) | ||
|
||
关于头梯度与链式法则 [菩提树下的杨过](http://www.cnblogs.com/yjmyzz/p/7783286.html)讲解的十分详细。 | ||
|
||
------------ | ||
|
||
### Welcome to Jupyter! | ||
|
||
- [Jupyter Try](https://jupyter.org/try) | ||
- [Jupyter Hub 教程](https://binderhub.readthedocs.io/en/latest/) | ||
|
||
This repo contains an introduction to [Jupyter](https://jupyter.org) and [IPython](https://ipython.org). | ||
|
||
You can also get this tutorial and run it on your laptop: | ||
|
||
git clone https://github.com/ipython/ipython-in-depth | ||
|
||
Install IPython and Jupyter: | ||
|
||
with [conda](https://www.anaconda.com/download): | ||
|
||
conda install ipython jupyter | ||
|
||
with pip: | ||
|
||
|
||
```sh | ||
# first, always upgrade pip! | ||
pip install --upgrade pip | ||
pip install --upgrade ipython jupyter | ||
``` | ||
Start the notebook in the tutorial directory: | ||
```sh | ||
cd ipython-in-depth | ||
jupyter notebook | ||
``` | ||
|
||
### Python 优秀教程 | ||
|
||
- [Learn Python](https://learnpython.org/) | ||
- [Learn Python the Hard Way](https://learnpythonthehardway.org/book/) | ||
- [Practical Machine Learning Tutorial with Python Introduction](https://pythonprogramming.net/machine-learning-tutorial-python-introduction/) 涵盖了与机器学习相关的概念、算法、理论、应用程序等. | ||
- [Introduction to Neural Networks](https://pythonprogramming.net/neural-networks-machine-learning-tutorial/) | ||
|
||
### 深度学习(Deep Learning,DL) | ||
|
||
虽然 Tensorflow 已经封装好了大部分 DL 的细节,但是如果不了解 DL 的相关核心概念,就会很难着手分析问题。首先强烈推荐教程[Neural Networks and Deep Learning](http://neuralnetworksanddeeplearning.com/),通读一遍,然后还有深度学习圣经[Deep Learning](http://www.deeplearningbook.org/),可以快速浏览或者只看自己不太明白的地方,还有[Welcome to the Deep Learning Tutorial! ](http://deeplearning.stanford.edu/tutorial/)分块介绍的,还有几篇blog,[Conv Nets: A Modular Perspective](https://colah.github.io/posts/2014-07-Conv-Nets-Modular/)讲的是卷积神经网络。图像识别用的比较多的就是卷积神经网络,这篇可以看看。 | ||
[Deep Learning Tutorials](http://deeplearning.net/tutorial/)也是不错的。 | ||
|
||
### TensorFlow 教程 | ||
|
||
- [官方教程](https://www.tensorflow.org/tutorials/) | ||
- [官方文档中文版](http://www.tensorfly.cn/) | ||
- [TensorFlow 文档中文翻译版](http://wiki.jikexueyuan.com/project/tensorflow-zh/) | ||
- [tensorflow cookbook](https://github.com/nfmcclure/tensorflow_cookbook) | ||
- [斯坦福教程](https://web.stanford.edu/class/cs20si/syllabus.html) | ||
- [Tensorflow的教程](http://learningtensorflow.com/) | ||
- [莫烦 TensorFlow](https://morvanzhou.github.io/tutorials/machine-learning/tensorflow/) | ||
- [Introduction to Deep Learning with TensorFlow](https://pythonprogramming.net/tensorflow-introduction-machine-learning-tutorial/) | ||
- [Deep Learning with TensorFlow - Creating the Neural Network Model](https://pythonprogramming.net/tensorflow-deep-neural-network-machine-learning-tutorial/) | ||
|
||
- [在Android端使用 TensorFlow](http://hp.stuhome.net/index.php/2017/02/17/tensorflow-on-android/) | ||
|
||
### 优质 blog | ||
|
||
- [Hackery, Math](http://acko.net/) | ||
- [colah's blog](https://colah.github.io/) | ||
- [ Café, bonne nuit](https://ireneli.eu/) | ||
- [Mike Bostock](https://bost.ocks.org/mike/) | ||
|
||
### 优秀课程资源 | ||
|
||
[使用TensorFlow |创建深度学习应用程序](https://www.kadenze.com/courses/creative-applications-of-deep-learning-with-tensorflow/info)以应用为导向的课程,通过鼓励探索创造性思维和深层神经网络的创造性应用,教你Tensorflow的基础知识以及最先进的算法,强烈鼓励尝试这门课程。这是唯一全面的在线课程,将教会你如何使用 Tensorflow 和开发您的创造潜力,了解如何应用这些技术创建神经网络。 | ||
|
||
建议你看一下[Deep Learning with TensorFlow](https://cognitiveclass.ai/courses/deep-learning-tensorflow/)课程。同样查看目录中的其他数据科学和机器学习课程。课程是完全免费的,并且许多都有完成证书和IBM支持的开放徽章。 | ||
|
||
### [如何开始学习使用TensorFlow?](https://yq.aliyun.com/articles/71257?utm_campaign=wenzhang&utm_medium=article&utm_source=QQ-qun&utm_content=m_12436) | ||
|
||
### 优秀的 GitHub | ||
|
||
- [ahangchen/GDLnotes](https://github.com/ahangchen/GDLnotes) | ||
- [ahangchen/GoogleML](https://github.com/ahangchen/GoogleML) | ||
- [wagamamaz](https://github.com/wagamamaz) | ||
|
||
### TensorFlow 学习[caicloud/tensorflow-tutorial](https://github.com/caicloud/tensorflow-tutorial) | ||
|
||
[TensorFlow 白皮书](https://www.jianshu.com/p/65dc64e4c81f) | ||
|
||
[git - the simple guide](https://rogerdudler.github.io/git-guide/) |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,4 @@ | ||
## 深度学习 | ||
|
||
- [eriklindernoren/PyTorch-YOLOv3](https://github.com/eriklindernoren/PyTorch-YOLOv3): 一个最小化的 YOLOv3 实现。 | ||
- [TuSimple/mx-maskrcnn](https://github.com/TuSimple/mx-maskrcnn): 一份 MXNet 实现的 Mask R-CNN。 |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -1,3 +1,4 @@ | ||
## 开源项目 | ||
|
||
- [embeddedartistry](https://github.com/embeddedartistry)/**[templates](https://github.com/embeddedartistry/templates)**: 一份开源项目的 README, CONTRIBUTING, GitHub templates 模板。 | ||
- [awesome-github-templates](https://github.com/devspace/awesome-github-templates): GitHub 的 Issues and Pull Requests 模板。 |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -1,9 +1,16 @@ | ||
## 推荐专栏 | ||
## 推荐 | ||
|
||
### 1. jekyll 推荐 | ||
|
||
1. Github Pages 的一份不错的主题模板,设置 `_config.yml` 为: | ||
|
||
```yml | ||
theme: minima | ||
``` | ||
- Github Pages 的一份不错的主题模板 minima ,设置 `_config.yml` 为:`theme: minima` 即可。 | ||
- [How-To-Ask-Questions-The-Smart-Way](https://github.com/ryanhanwu/How-To-Ask-Questions-The-Smart-Way): 学习如何提问? | ||
- [pytorch_geometric](https://github.com/rusty1s/pytorch_geometric): 一份由 Pytorch 编写的图神经网络库。 | ||
- [maskrcnn-benchmark](https://github.com/facebookresearch/maskrcnn-benchmark): 一份 Facebook 开源的基准(benchmark)算法工程,其中包含检测、分割和人体关键点等算法。 | ||
- [fastai](https://github.com/fastai/fastai): 一个课程平台,一个讨论社区,也是一个PyTorc的顶层框架。Fastai的理念就是让神经网络没那么望而生畏,其课程也是采用项目驱动的方式教学。(不同于很多深度学习课程先教授大量的数学基础,各有利弊 | ||
- [深度好文 |Matplotlib 可视化最有价值的 50 个图表(附完整 Python 源代码)](https://zhuanlan.zhihu.com/p/53308606) | ||
- [Python Echarts Plotting Library](https://github.com/pyecharts/pyecharts): 一个由百度开源的数据可视化,凭借着良好的交互性,精巧的图表设计,得到了众多开发者的认可。而 Python 是一门富有表达力的语言,很适合用于数据处理。当数据分析遇上数据可视化时,pyecharts 诞生了。 | ||
- [homemade-machine-learning](https://github.com/trekhleb/homemade-machine-learning): 一个 Python 实现的机器学习库。 | ||
- [selenium-document](https://github.com/xuyichenmo/selenium-document): 一份 Selenium 的中文文档。 | ||
- [gluon-face](https://github.com/THUFutureLab/gluon-face): 一份由 Gluon 实现的人脸识别库。 | ||
- [TuSimple/simpledet](https://github.com/TuSimple/simpledet)|[知乎](https://zhuanlan.zhihu.com/p/55972055): SimpleDet: 一套简单通用的目标检测与物体识别框架。 | ||
- [deepinsight/insightface](https://github.com/deepinsight/insightface): InsightFace: 2D and 3D Face Analysis Project。 | ||
- [awesome-computer-vision](https://github.com/jbhuang0604/awesome-computer-vision): A curated list of awesome computer vision resources. | ||
- [open-mmlab/mmcv](https://github.com/open-mmlab/mmcv): mmcv 是一个基础库,主要分为两个部分,一部分是和 deep learning framework 无关的一些工具函数,比如 IO/Image/Video 相关的一些操作,另一部分是为 PyTorch 写的一套训练工具,可以大大减少用户需要写的代码量,同时让整个流程的定制变得容易。 |