1 Keras
1.1 一个基本示例
import numpy as np
from keras.models import Sequential
from keras.layers import Dense
# 1.加载数据集
data = np.random.random((1000, 100)) # 创建样本
labels = np.random.randint(2, size=(1000, 1)) # 创建只有0，1两类的标签
# 2.构建模型
model = Sequential() # 构建序列结构
model.add(Dense(32, activation='relu', input_dim=100)) # 往序列结构中添加拥有32个神经元的全连接层，输入是100维向量（注意默认忽略批量维度）
model.add(Dense(1, activation='sigmoid')) # 往序列结构中添加拥有1个神经元的全连接层
# 3.编译模型
model.compile(optimizer='rmsprop',
loss='binary_crossentropy',
metrics=['accuracy'])
# 4.训练模型
model.fit(data, labels, epochs=10, batch_size=32)
# 5.预测模型
predictions = model.predict(data)
2 数据
2.1 Keras数据设置
from keras.datasets import boston_housing, mnist, cifar10, imdb
(x_train1, y_train1), (x_test1, y_test1) = mnist.load_data()
(x_train2, y_train2), (x_test2, y_test2) = boston_housing.load_data()
(x_train3, y_train3), (x_test3, y_test3) = cifar10.load_data()
(x_train4, y_train4), (x_test4, y_test4) = imdb.load_data(num_words=20000)
num_classes = 10
3 模型结构
3.1 Sequential模型
from keras.models import Sequential
model = Sequential()
model2 = Sequential()
model3 = Sequential()
3.2 多层感知器（MLP）
3.2.1 二元分类
from keras.layers import Dense
model.add(Dense(12, input_dim=8, kernel_initializer='uniform', activation='relu'))
model.add(Dense(8, kernel_initializer='uniform', activation='relu'))
model.add(Dense(1, kernel_initializer='uniform', activation='sigmoid'))
3.2.2 多类别分类
from keras.layers import Dense, Dropout
model.add(Dense(512, activation='relu', input_shape=(784,)))
model.add(Dropout(0.2))
model.add(Dense(512, activation='relu'))
model.add(Dropout(0.2))
model.add(Dense(10, activation='softmax'))
3.2.3 回归
from keras.layers import Dense
model.add(Dense(64, activation='relu', input_dim=train_data.shape[1]))
model.add(Dense(1))
3.3 卷积神经网络（CNN）
from keras.layers import Activation,Conv2D,MaxPooling2D,Flatten
model.add(Conv2D(32, (3,3), padding='same', input_shape=x_train.shape[1:]))
model.add(Activation('relu'))
model.add(Conv2D(32, (3,3)))
model.add(Activation('relu'))
model.add(MaxPooling2D(pool_size=(2,2)))
model.add(Dropout(0.25))
model.add(Conv2D(64, (3,3), padding='same'))
model.add(Activation('relu'))
model.add(Conv2D(64, (3, 3)))
model.add(Activation('relu'))
model.add(MaxPooling2D(pool_size=(2, 2)))
model.add(Dropout(0.25))
model.add(Flatten())
model.add(Dense(512))
model.add(Activation('relu'))
model.add(Dropout(0.5))
model.add(Dense(num_classes))
model.add(Activation('softmax'))
3.4 循环神经网络（RNN）
from keras.layers import Embedding,LSTM
model.add(Embedding(20000,128))
model.add(LSTM(128,dropout=0.2,recurrent_dropout=0.2))
model.add(Dense(1,activation='sigmoid'))
4 预处理
4.1 序列填充
将数据填充至指定长度(maxlen)，默认填充值value = 0.0
from keras.preprocessing import sequence
x_train = sequence.pad_sequences(x_train4,maxlen=80)
x_test = sequence.pad_sequences(x_test4,maxlen=80)
4.2 创建虚拟变量
from keras.utils import to_categorical
Y_train = to_categorical(y_train, num_classes)
Y_test = to_categorical(y_test, num_classes)
Y_train3 = to_categorical(y_train3, num_classes)
Y_test3 = to_categorical(y_test3, num_classes)
4.3 训练集、测试集分离
from sklearn.model_selection import train_test_split
X_train,X_test,y_train,y_test = train_test_split(X,
y,
test_size=0.33,
random_state=42)
4.4 标准化/归一化
from sklearn.preprocessing import StandardScaler
scaler = StandardScaler().fit(x_train)
standardized_X = scaler.transform(x_train)
standardized_X_test = scaler.transform(x_test)
5 模型细节提取
5.1 模型输出形状
model.output_shape
5.2 模型总结
model.summary()
5.3 get模型参数
model.get_config()
5.4 get神经网络weights
model.get_weights()
6 编译模型
6.1 多层感知器
6.1.1 二分类问题
model.compile(optimizer='adam',
loss='binary_crossentropy',
metrics=['accuracy'])
6.1.2 多分类问题
model.compile(optimizer='rmsprop',
loss='categorical_crossentropy',
metrics=['accuracy'])
6.1.3 回归问题
model.compile(optimizer='rmsprop',
loss='mse',
metrics=['mae'])
6.2 循环神经网络
model.compile(loss='binary_crossentropy',
optimizer='adam',
metrics=['accuracy'])
7 训练模型
model.fit(x_train4,
y_train4,
batch_size=32,
epochs=15,
verbose=1,
validation_data=(x_test4, y_test4))
8 评价模型
score = model.evaluate(x_test,
y_test,
batch_size=32)
9 预测结果
model.predict(x_test4, batch_size=32)
model.predict_classes(x_test4,batch_size=32)
10 保存/载入模型
from keras.models import load_model
model.save('model_file.h5')
my_model = load_model('my_model.h5')
11 模型参数调节
11.1 优化参数
from keras.optimizers import RMSprop
opt = RMSprop(lr=0.0001, decay=1e-6)
model.compile(loss='categorical_crossentropy',
optimizer=opt,
metrics=['accuracy'])
11.2 模型提前终止
from keras.callbacks import EarlyStopping
early_stopping_monitor = EarlyStopping(patience=2)
model.fit(x_train4,
y_train4,
batch_size=32,
epochs=15,
validation_data=(x_test4,y_test4),
callbacks=[early_stopping_monitor])