tf2_emnist_test/model.py at master · JiJiFlyer/tf2_emnist_test · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
import tensorflow as tf
from tensorflow import keras
from tensorflow.keras import layers, models

def LeNet_inference(in_shape, outputclass, dropout):
    model = models.Sequential(name='LeNet')
    # 第1层卷积，卷积核大小为3*3，32个，28*28为待训练图片的大小
    model.add(layers.Conv2D(
        32, (3, 3), padding='same', activation='relu', input_shape=(in_shape[0],in_shape[1],in_shape[2])))
    model.add(layers.MaxPooling2D((2, 2), padding='same'))
    # 第2层卷积，卷积核大小为3*3，64个
    model.add(layers.Conv2D(64, (3, 3), activation='relu', padding='same'))
    model.add(layers.MaxPooling2D((2, 2), padding='same'))
    # 第3层卷积，卷积核大小为3*3，64个
    model.add(layers.Conv2D(64, (3, 3), activation='relu', padding='same'))

    model.add(layers.Flatten())
    model.add(layers.Dense(80, activation='relu'))
    model.add(layers.Dense(outputclass, activation='softmax'))

    model.summary()
    #25s 13ms/step - loss: 0.3786 - accuracy: 0.8615 - val_loss: 0.4692 - val_accuracy: 0.8417
    model.compile(
        loss='sparse_categorical_crossentropy',
        optimizer=tf.keras.optimizers.Adam(dropout),
        metrics=['accuracy'],
    )

    return model


def AlexNet_inference(in_shape, outputclass, dropout):
    model = models.Sequential(name='AlexNet')

    # model.add(layers.Conv2D(96,(11,11),strides=(4,4),input_shape=(in_shape[1],in_shape[2],in_shape[3]),
                # padding='same',activation='relu',kernel_initializer='uniform'))

    model.add(layers.Conv2D(96,(11,11),strides=(2,2),input_shape=(in_shape[0],in_shape[1],in_shape[2]),
                padding='same',activation='relu',kernel_initializer='uniform'))
    model.add(layers.MaxPooling2D(pool_size=(3,3),strides=(2,2)))
    model.add(layers.Conv2D(256,(5,5),strides=(1,1),padding='same',activation='relu',kernel_initializer='uniform'))
    model.add(layers.MaxPooling2D(pool_size=(3,3),strides=(2,2)))
    model.add(layers.Conv2D(384,(3,3),strides=(1,1),padding='same',activation='relu',kernel_initializer='uniform'))
    model.add(layers.Conv2D(384,(3,3),strides=(1,1),padding='same',activation='relu',kernel_initializer='uniform'))
    model.add(layers.Conv2D(256,(3,3),strides=(1,1),padding='same',activation='relu',kernel_initializer='uniform'))
    model.add(layers.MaxPooling2D(pool_size=(2,2),strides=(2,2)))
    model.add(layers.Flatten())
    model.add(layers.Dense(2048,activation='relu'))
    model.add(layers.Dropout(dropout))
    model.add(layers.Dense(2048,activation='relu'))
    model.add(layers.Dropout(dropout))
    model.add(layers.Dense(outputclass,activation='softmax'))
    model.compile(optimizer=keras.optimizers.Adam(),
                loss='sparse_categorical_crossentropy', #不能直接用函数，否则在与测试加载模型不成功！
                metrics=['accuracy'])
    model.summary()
    # 75s 38ms/step - loss: 0.6148 - accuracy: 0.8204 - val_loss: 0.5183 - val_accuracy: 0.8379
    return model

def VGG16_inference(in_shape, outputclass, dropout):
    network_VGG16 = Sequential([
        # 第一层
        layers.Conv2D(64, kernel_size=3, strides=1, padding='same', activation='relu'),
        # 第二层
        layers.Conv2D(64, kernel_size=3, strides=1, padding='same', activation='relu'),
        layers.MaxPooling2D(pool_size=2, strides=2),
        # 第三层
        layers.Conv2D(128, kernel_size=3, strides=1, padding='same', activation='relu'),
        # 第四层
        layers.Conv2D(128, kernel_size=3, strides=1, padding='same', activation='relu'),
        layers.MaxPooling2D(pool_size=2, strides=2),
        # 第五层
        layers.Conv2D(256, kernel_size=3, strides=1, padding='same', activation='relu'),
        # 第六层
        layers.Conv2D(256, kernel_size=3, strides=1, padding='same', activation='relu'),
        # 第七层（新增卷积层1*1*256）
        layers.Conv2D(256, kernel_size=1, strides=1, padding='same', activation='relu'),
        layers.MaxPooling2D(pool_size=2, strides=2),
        # 第八层
        layers.Conv2D(512, kernel_size=3, strides=1, padding='same', activation='relu'),
        # 第九层
        layers.Conv2D(512, kernel_size=3, strides=1, padding='same', activation='relu'),
        # 第十层（新增卷积层1*1*512）
        layers.Conv2D(512, kernel_size=1, strides=1, padding='same', activation='relu'),
        layers.MaxPooling2D(pool_size=2, strides=2),
        # 第十一层
        layers.Conv2D(512, kernel_size=3, strides=1, padding='same', activation='relu'),
        # 第十二层
        layers.Conv2D(512, kernel_size=3, strides=1, padding='same', activation='relu'),
        # 第十三层（新增卷积层1*1*512）
        layers.Conv2D(512, kernel_size=1, strides=1, padding='same', activation='relu'),
        layers.MaxPooling2D(pool_size=2, strides=2),
        layers.Flatten(),  # 拉直 7*7*512
        # 第十四层
        layers.Dense(1024, activation='relu'),
        layers.Dropout(rate=dropout),
        # 第十五层
        layers.Dense(128, activation='relu'),
        layers.Dropout(rate=dropout),
        # 第十六层
        layers.Dense(5, activation='softmax')
    ])
    network_VGG16.compile(optimizer=keras.optimizers.Adam(),
            loss='sparse_categorical_crossentropy', #不能直接用函数，否则在与测试加载模型不成功！
            metrics=['accuracy'])
    network_VGG16.summary()  # 打印各层参数表
    return network_VGG16

#--------------------------#
# RESNET
#--------------------------#
def res_net_block(input_data, filters, conv_size):
  # CNN层
  x = layers.Conv2D(filters, conv_size, activation='relu', padding='same')(input_data)
  x = layers.BatchNormalization()(x)
  x = layers.Conv2D(filters, conv_size, activation=None, padding='same')(x)
  # 第二层没有激活函数
  x = layers.BatchNormalization()(x)
  # 两个张量相加
  x = layers.Add()([x, input_data])
  # 对相加的结果使用ReLU激活
  x = layers.Activation('relu')(x)
  # 返回结果
  return x

def ResNet_inference(input_shape, n_classes, dropout):
    inputs = keras.Input(shape=input_shape)
    x = layers.Conv2D(32, 3, activation='relu')(inputs)
    x = layers.Conv2D(64, 3, activation='relu')(x)
    x = layers.MaxPooling2D(3)(x)
    num_res_net_blocks = 10
    for i in range(num_res_net_blocks):
        x = res_net_block(x, 64, 3)
    # 添加一个CNN层
    x = layers.Conv2D(64, 3, activation='relu')(x)
    # 全局平均池化GAP层
    x = layers.GlobalAveragePooling2D()(x)
    # 几个密集分类层
    x = layers.Dense(256, activation='relu')(x)
    # 退出层
    x = layers.Dropout(dropout)(x)
    outputs = layers.Dense(n_classes, activation='softmax')(x)
    res_net_model = keras.Model(inputs, outputs)
    res_net_model.compile(optimizer=keras.optimizers.Adam(),
              loss='sparse_categorical_crossentropy',
              metrics=['accuracy'])
    res_net_model.summary()
    #  105s 53ms/step - loss: 0.2584 - accuracy: 0.8978 - val_loss: 0.3838 - val_accuracy: 0.8743
    #  2000*20steps开始过拟合
    return res_net_model