用keras实现cnn分类

搞了挺久，包括正确率的提高，还有各种错误之后好了

import tensorflow
import keras
from keras.layers import Dense, Dropout, Activation, Flatten
from keras.layers.normalization import BatchNormalization
from keras.models import Sequential
from keras.layers import Conv2D, MaxPooling2D
from keras.utils import np_utils
import numpy as np 


#二维卷积层Conv2D
#keras.layers.convolutional.Conv2D(filters, kernal_size, strides = (1,1), padding = 'valdid')
#fileters是卷积核的数目，kernal_size是卷积核的尺寸，strides是卷积核移动的步长，padding是边界模式（一般分为same和valid两种，same是全填充，边界补零）
#二维池化层Maxpooling，池化层：对输入的特征图进行压缩，一方面使特征图变小，简化网络计算复杂度；一方面进行特征压缩，提取主要特征
#keras.layers.pooling.Maxpooling2D(pool_size = (2,2), strides = None, padding = 'valid')
#Activation层：keras.layers.core.Activation(activation),激活函数可以为：softmax, softplus, softsign, relu, sigmoid, hard_sigmoid, linear等
#Droupout层：kera