1.time.time()/clock()
https://www.cnblogs.com/bettermanlu/archive/2011/09/19/2181529.html
前者挂钟时间，后者处理器CPU时间。
挂钟时间也称为经过时间或运行时间。 与CPU时间相比，挂钟时间通常更长，因为执行测量程序的CPU也可能同时执行其他程序的指令。
2.reload(sys)
https://blog.csdn.net/qq_36711420/article/details/79382327
主要进行编码格式的转换
3.local variable 'xxx' referenced before assignment
http://blog.sina.com.cn/s/blog_4b9eab320100q2l5.html
4. 8大排序算法
5.希尔排序
https://blog.csdn.net/weixin_37818081/article/details/79202115
6.归并排序
https://blog.csdn.net/k_koris/article/details/80508543
7.python中的Swap函数
https://blog.csdn.net/hyqsong/article/details/47864753
8.CNN
https://nndl.github.io/
9.LeNet-5学习及简单实现
https://www.charleychai.com/blogs/2018/ai/NN/lenet.html
https://blog.csdn.net/qq_15192373/article/details/78536107
11.CNN中补齐same/valid两种方式
https://blog.csdn.net/wuzqChom/article/details/74785643
https://blog.csdn.net/qq_30979017/article/details/79407720
import keras def test_model(input): input=keras.layers.Input(input) x=keras.layers.Conv2D(32,(5,5),strides=(2,2),padding='same')(input) return keras.models.Model(input,x) model=test_model((7,7,3))
12.conv1D和conv2D的区别
https://blog.csdn.net/hahajinbu/article/details/79535172
//get
13.instance函数
https://www.runoob.com/python/python-func-isinstance.html
14.只有一个元素的元组加逗号
https://blog.csdn.net/carrey_0612/article/details/79932265
16.ndarray的多维理解
https://zhuanlan.zhihu.com/p/39287693
17.读取csv文件
https://www.cnblogs.com/cloud-ken/p/8432999.html
18.ValueError: sequence too large; cannot be greater than 32
https://stackoverflow.com/questions/17688094/numpy-array-sequence-too-large
https://stackoverflow.com/questions/39325930/numpy-ndarray-with-more-that-32-dimensions
19.DataFrame的切片操作
https://blog.csdn.net/LY_ysys629/article/details/55224284
20.np_utils.to_categorical
21.对数组矩阵补齐padding
https://blog.csdn.net/zenghaitao0128/article/details/78713663
23.Keras可视化训练误差和验证误差
https://machinelearningmastery.com/display-deep-learning-model-training-history-in-keras/
24.flatten展平的作用
https://blog.csdn.net/program_developer/article/details/80853425
25.Keras实现LeNet-5/tf实现（好文）
https://medium.com/@mgazar/lenet-5-in-9-lines-of-code-using-keras-ac99294c8086
27.one-hot与数组相互转换
https://blog.csdn.net/ChaoFeiLi/article/details/89363117
28.递归和循环的比较
http://landcareweb.com/questions/791/di-gui-bi-xun-huan-geng-kuai-ma
29.函数式编程语言
http://www.ruanyifeng.com/blog/2012/04/functional_programming.html
30.Keras之ImageDataGenerator()
https://www.jianshu.com/p/d23b5994db64
31.numpy.expand_dims
https://blog.csdn.net/qq_16949707/article/details/53418912
32.DN中的过拟合问题及解决办法
https://www.jianshu.com/p/86051c14d434
35.标准差的计算
https://www.cnblogs.com/webRobot/p/7722820.html
为何分母为n-1
https://blog.csdn.net/Hearthougan/article/details/77859173
36.playground教程
https://www.jianshu.com/p/5f83defc7615
37.二分查找时间复杂度
https://blog.csdn.net/frances_han/article/details/6458067
38.py中的二维数组申请
https://www.cnblogs.com/btchenguang/archive/2012/01/30/2332479.html
39.牛顿法-优化算法
https://blog.csdn.net/google19890102/article/details/41087931
40.最优化算法overview
http://ruder.io/optimizing-gradient-descent/
41.shel脚本编程入门
https://www.runoob.com/linux/linux-shell-include-file.html
https://lg1024.com/post/shell_01.html
42._import_动态加载模块
43.打开pkl文件
https://blog.csdn.net/NewstarSouth/article/details/47092037
44.io.open和open
https://stackoverflow.com/questions/33891373/difference-between-io-open-vs-open-in-python
46.tf的GPU和CPU的区别
https://blog.csdn.net/huangx06/article/details/78835360
47.全角半角转换
https://www.cnblogs.com/kaituorensheng/p/3554571.htm
48.2.7中的zip和izip
https://blog.csdn.net/orangleliu/article/details/23737701
>>> a=[[0, 1, 2], [3, 4, 5], [6, 7, 8]] >>> b=[[0, 1, 2], [3, 4, 5], [6, 7, 8]] >>> d=list(zip(a,b)) >>> d [([0, 1, 2], [0, 1, 2]), ([3, 4, 5], [3, 4, 5]), ([6, 7, 8], [6, 7, 8])]#第一个元祖包含a/b的第一个元素（list类型）
49.linux中查看文件编码格式
https://jingyan.baidu.com/article/36d6ed1f6fc8b71bcf48838e.html
50.tf中的参数初始化方法