【无标题】keras+mnist+LeNet_5实现

keras+mnist+LeNet_5实现
具体实现步骤：

1. 加载Mnist数据集，可视化训练集样本数据
2. 对数据集进行标准化预处理（归一化和独热编码处理）
3. keras构建LeNet_5经典模型，打印模型参数
4. 模型反向编译，模型训练
5. 模型训练过程acc和loss可视化展示

import keras
from keras import datasets
from  keras.layers import Conv2D,MaxPool2D,Dense,Flatten,Activation
from  keras.models import  Sequential
from keras.optimizers import  SGD
import numpy as np
from matplotlib import pyplot as plt
# 1. 数据集处理
# 加载数据，输出样本和标签数据
(X_train,Y_trian),(X_test,Y_test) = datasets.mnist.load_data()
print("训练集：",X_train.shape)
print("训练集标签",Y_trian.shape)
print("测试集",X_test.shape)
print("测试集标签",Y_test.shape)

# 2可视化图像
plt.figure(figsize=(15,15))
plt.title('mnist')
for i in range (20):
    plt.subplot(4, 5, i + 1)
    plt.xticks([])
    plt.yticks([])
    plt.grid(False)
    plt.imshow(X_test[i])
    plt.xlabel(Y_test[i])
    plt.savefig('../TermimalTest2020/Images/MnistDate.jpg')
plt.show()

# 手写体种类，0~9 一共10个类别
num_class = 10
# 3图像数据标准化处理
X_train = np.reshape(X_train,[-1,28,28,1])
X_test  = np.reshape(X_test ,[-1,28,28,1])

X_train = X_train/255.0
X_test = X_test/255.0
# 独热编码处理
Y_trian = keras.utils.np_utils.to_categorical(Y_trian,num_classes=num_class)
Y_test = keras.utils.np_utils.to_categorical(Y_test,num_classes=num_class)
print(Y_trian[0])


# 4. 模型LeNET-5模型
model = Sequential([
    keras.layers.Conv2D(6, (5,5), activation='tanh', input_shape=(28,28,1),padding='same'),
    keras.layers.MaxPool2D((2,2),strides=(1,1)),
    keras.layers.Conv2D(16, (5,5), activation='tanh',padding='same'),
    keras.layers.MaxPool2D((2,2),strides=(1,1)),
    keras.layers.Flatten(),
    keras.layers.Dense(120, activation='tanh'),
    keras.layers.Dense(84, activation='tanh'),
    keras.layers.Dense(num_class),
    keras.layers.Activation('softmax')
])
model.summary()
# 5. 数据反向传播
model.compile(loss='categorical_crossentropy',
             optimizer=SGD(lr=0.01, decay=1e-6, momentum=0.9, nesterov=True),
             metrics=['acc'])
#保存权值模型
model.save_weights('../models/mnist.h5')
# 6. 数据训练
epochs = 10
batch_size =32
result=model.fit(x=X_train,y=Y_trian,epochs=epochs,batch_size=batch_size,verbose=2,
    validation_split=0.2)
# 模型训练可视化
plt.subplot(1, 2, 1)
plt.plot(range(1, epochs + 1), result.history["loss"])
plt.plot(range(1, epochs + 1), result.history["val_loss"])
plt.title("loss")
plt.xlabel("epochs")
plt.legend()
plt.ylabel("val/train loss")

plt.subplot(1, 2, 2)
plt.plot(range(1, epochs + 1), result.history["acc"])
plt.plot(range(1, epochs + 1), result.history["val_acc"])
plt.title("acc")
plt.xlabel("epochs")
plt.ylabel("val/train acc")
plt.legend()
plt.show()
plt.close()