keras 的 example 文件 cifar10_cnn.py 解析

这个示例很简单，就是从cifar10中读取数据集，通过卷积神经网络进行图像识别

输入数据的shape

x_train.shape (50000, 32, 32, 3)
y_train.shape (50000, 10)

 

神经网络结构：

________________________________________________________________________________
Layer (type)                        Output Shape                    Param #
================================================================================
conv2d_1 (Conv2D)                   (None, 32, 32, 32)              896
________________________________________________________________________________
activation_1 (Activation)           (None, 32, 32, 32)              0
________________________________________________________________________________
conv2d_2 (Conv2D)                   (None, 30, 30, 32)              9248
________________________________________________________________________________
activation_2 (Activation)           (None, 30, 30, 32)              0
________________________________________________________________________________
max_pooling2d_1 (MaxPooling2D)      (None, 15, 15, 32)              0
________________________________________________________________________________
dropout_1 (Dropout)                 (None, 15, 15, 32)              0
________________________________________________________________________________
conv2d_3 (Conv2D)                   (None, 15, 15, 64)              18496
________________________________________________________________________________
activation_3 (Activation)           (None, 15, 15, 64)              0
________________________________________________________________________________
conv2d_4 (Conv2D)                   (None, 13, 13, 64)              36928
________________________________________________________________________________
activation_4 (Activation)           (None, 13, 13, 64)              0
________________________________________________________________________________
max_pooling2d_2 (MaxPooling2D)      (None, 6, 6, 64)                0
________________________________________________________________________________
dropout_2 (Dropout)                 (None, 6, 6, 64)                0
________________________________________________________________________________
flatten_1 (Flatten)                 (None, 2304)                    0
________________________________________________________________________________
dense_1 (Dense)                     (None, 512)                     1180160
________________________________________________________________________________
activation_5 (Activation)           (None, 512)                     0
________________________________________________________________________________
dropout_3 (Dropout)                 (None, 512)                     0
________________________________________________________________________________
dense_2 (Dense)                     (None, 10)                      5130
________________________________________________________________________________
activation_6 (Activation)           (None, 10)                      0
================================================================================
Total params: 1,250,858
Trainable params: 1,250,858
Non-trainable params: 0
________________________________________________________________________________

代码同时演示了 ImageDataGenerator 的使用

 

——————————————————————

总目录

keras的example文件解析