基本去噪自编码器的实现

import tensorflow as tf import matplotlib.pyplot as plt import numpy as np

查看tensorflow版本

1	tf.__version__

导入mnist数据集

1	(x_train,_),(x_test,_) = tf.keras.datasets.mnist.load_data()

查看数据维度

1	x_train.shape,x_test.shape

对数据进行reshape 28*28->784

1 2	x_train = x_train.reshape(x_train.shape[0], -1) x_test = x_test.reshape(x_test.shape[0], -1)

数据归一化

1 2	x_train = tf.cast(x_train,tf.float32)/255 x_test = tf.cast(x_test,tf.float32)/255

噪声的添加

1
2
3

factor = 0.5 #增加噪声多少
x_train_noise = x_train + factor*np.random.normal(loc=0.0,scale=1.0,size=x_train.shape)
x_test_noise = x_test + factor*np.random.normal(loc=0.0,scale=1.0,size=x_test.shape)

数据范围规定

1 2	x_train_noise = np.clip(x_train_noise,0.,1.)#加噪后数值可能会不在0~1之间将范围规定到之中，大于1变成1小于0变成00 x_test_noise = np.clip(x_test_noise,0.,1.)

查看加噪后的图片

n=10
plt.figure(figsize=(10,2))
for i in range(1,n):
    ax = plt.subplot(1,n,i)
    plt.imshow(x_train_noise[i].reshape(28,28))

构建模型

input_size = 784
hidden_size = 32
output_size = 784

input = tf.keras.layers.Input(shape=(input_size,))

# Encoder
en = tf.keras.layers.Dense(hidden_size, activation='relu')(input)

# Decoder
de = tf.keras.layers.Dense(output_size, activation='sigmoid')(en)

model = tf.keras.Model(inputs=input, outputs=de)

查看模型数据流动

1	tf.keras.utils.plot_model(model,show_shapes=True)#可视化

模型配置

1
2
3

model.compile(optimizer=tf.keras.optimizers.Adam(0.001),
              loss='mse' 
)

模型训练

model.fit(x_train_noise, x_train,
                epochs=50,
                batch_size=256,
                shuffle=True,
                validation_data=(x_test_noise,x_test)
)

模型测试

编码

1	encode = tf.keras.Model(inputs=input,outputs=en)

model.layers[-1].name
1
'dense_3'

解码

1
2
3

input_de = tf.keras.layers.Input(shape=(hidden_size,))
output = model.layers[-1](input_de)
decode = tf.keras.Model(inputs=input_de, outputs=output)

输出

encode_test = encode(x_test_noise)#TensorShape([10000, 32])
de_test = decode.predict(encode_test)#(10000, 784)
n = 10
x_test = x_test.numpy()
plt.figure(figsize=(20,4))
for i in range(1,n):
    ax = plt.subplot(2,n,i)
    plt.imshow(x_test_noise[i].reshape(28,28))
    ax = plt.subplot(2,n,i+n)
    plt.imshow(de_test[i].reshape(28,28))