用Tensorflow API:tf.keras搭建网络八股
六步法
import
train,test
model=tf.keras.models.Sequential
model.compile
model.fit
model.summary:可以打印出网络的结构和参数统计
model=tf.keras.models.Sequential([网络结构]) #表述各层网络
#拉直层
tf.keras.layers.Flatten()
#全连接层
tf.keras.layer.Dense(神经元个数,activation='激活函数',kernel_regularizer=正则化)
#activation(字符串给出):relu,softmax,sigmoid,tanh
#kernel_regularizer:tf.keras.regularizers.l1(),tf.keras.regularizers.l2()
#卷积层
tf.keras.layers.Conv2D(filters=卷积核个数,kernwl_size=卷积核大小,strides=卷积步长,padding='valid'or'same')
#LSTM层
tf.keras.layer>LSTM()
model.compile(optimizer=优化器,loss=损失函数,metrics=['准确率']
#optimizer可选
'sgd' or tf.keras.optimizer.SGD(lr=学习率,momentum=动量参数)
'adagrad' or tf.keras.optimizer.Adagrad(lr=学习率)
'adadelts' or tf.keras.optimizers.Adadelta(lr=学习率)
'adam' or tf.keras.optimizers.Adam(lr=学习率,beta_1=0.9,beta-2=0.999)
#loss可选
'mse' or tf.keras.losses.MeanSquaredError()
'sparse_categorical_crossentropy' or tf.keras.losses.SparseCategoricalCrossentropy(from_logits=False) #是否是原始输出,有无经概率分布的输出
#metrics可选
'accuracy':y_和y都是数值,如y_=[1],y=[1]
'categorical_accuracy': y_和y都是独热编码(概率分布),如y_=[0,1,0],y=[0.256,0.695,0.048]
'sparse_categorical_accuracy':y_是数值,y是独热编码(概率分布),如y_=[1],y=[0.256,0.695,0.048]
model.fit(训练集的输入特征,训练集的标签,batch_size=batch大小,epochs=循环次数,
validation_data=(测试集的输入特征,测试集的标签),
validation_split=从训练集划分多少比例给测试集,
validation_freq=多少次epoch测试一次)
示例
# todo 1.导包
import tensorflow as tf
from sklearn import datasets
import numpy as np
# todo 2.获取数据
x_train = datasets.load_iris().data
y_train = datasets.load_iris().target
# 数据乱序
np.random.seed(116)
np.random.shuffle(x_train)
np.random.seed(116)
np.random.shuffle(y_train)
np.random.seed(116)
# todo 3.描述网络结构
model = tf.keras.models.Sequential([
tf.keras.layers.Dense(3, activation='softmax', kernel_regularizer=tf.keras.regularizers.l2())
])
# todo 4.优化器,损失函数
model.compile(optimizer=tf.keras.optimizers.SGD(lr=0.1),
loss=tf.keras.losses.SparseCategoricalCrossentropy(from_logits=False),
metrics=['sparse_categorical_accuracy']
)
# todo 5.训练
model.fit(
x_train, y_train, batch_size=32, epochs=500, validation_split=0.2, validation_freq=20
)
# todo 6.打印网格结构和参数统计
model.summary()