22-模型的保存与加载

api

api

模型的保存是保存为checkpoint文件。summary保存的文件为events文件

import tensorflow as tf
def myregression():
    """
    自实现一个线性回归预测
    :return: None
    """
    with tf.variable_scope("variable"):
        #准备数据
        x = tf.random_normal([100, 1], mean=1.75, stddev=0.5, name="x_data")
        y_true = tf.matmul(x, [[0.7]]) + 0.8 #矩阵相乘必须是2维的
    with tf.variable_scope("model"):
        #建立线回归模型
        weight = tf.Variable(tf.random_normal([1, 1], mean=0.0, stddev=1.0, name="weight"))
        bias = tf.Variable(0.0, name="bias")
        y_predict = tf.matmul(x, weight) + bias
    with tf.variable_scope("loss"):
        #建立损失函数，均方误差
        loss = tf.reduce_mean(tf.square(y_predict-y_true)) #reduce_mean是计算平均值
    with tf.variable_scope("optimizer"):
        #梯度下降优化损失
        train_op = tf.train.GradientDescentOptimizer(learning_rate=0.1).minimize(loss) #梯度下降去进行优化，即最小化损失，所以后面加了minimize

    #1. 收集变量：一般在会话之前
    tf.summary.scalar("losses", loss)
    tf.summary.histogram("weights", weight)

    #2. 合并变量，写入事件文件
    #定义合并变量的op
    merged = tf.summary.merge_all()

    #定义一个初始化变量的op
    init_op = tf.global_variables_initializer()

    #定义一个保存模型的实例
    saver = tf.train.Saver()

    #通过会话运行程序
    with tf.Session() as sess:
        #初始化变量
        sess.run(init_op)
        #打印随机初始化的权重和偏置值
        print("随机初始化的参数权重为：\n", weight.eval(), "\n偏置为：\n", bias.eval())
        #运行优化
        #循环训练优化
        filewriter = tf.summary.FileWriter("./", graph=sess.graph)
        for i in range(1000):
            sess.run(train_op)
            print("优化",i,"次优化过后的参数权重为：", weight.eval(), " 偏置为：", bias.eval())
            #运行合并的tensor
            summary = sess.run(merged)
            #把每次的值写入文件
            filewriter.add_summary(summary, i)
            if i % 100 == 0:
                saver.save(sess, "./model") #model是文件名
    return None
import os
def restoremodel():
    """
    加载模型
    :return:None
    """
    with tf.variable_scope("variable"):
        x = tf.random_normal([100, 1], mean=1.75, stddev=0.5, name="x_data")
        y_true = tf.matmul(x, [[0.7]]) + 0.8  # 矩阵相乘必须是2维的
    with tf.variable_scope("model"):
        weight = tf.Variable(tf.random_normal([1, 1], mean=0.0, stddev=1.0, name="weight"))
        bias = tf.Variable(0.0, name="bias")
        y_predict = tf.matmul(x, weight) + bias
    with tf.variable_scope("loss"):
        loss = tf.reduce_mean(tf.square(y_predict - y_true))  # reduce_mean是计算平均值
    with tf.variable_scope("optimizer"):
        train_op = tf.train.GradientDescentOptimizer(learning_rate=0.1).minimize(
            loss)  # 梯度下降去进行优化，即最小化损失，所以后面加了minimize
    # init_op = tf.global_variables_initializer()
    """注意：在恢复模型的时候，就不能再初始化所有变量"""
    saver = tf. train.Saver()
    with tf.Session() as sess:
        if os.path.exists("./checkpoint"):
            #加载模型，覆盖之前的参数
            saver.restore(sess, "./model")  #文件名即可，不需要加后缀。这里的文件名即为model
        # sess.run(init_op)
        for i in range(500):
            # sess.run(train_op)
            print("优化",i,"次优化过后的参数权重为：", weight.eval(), " 偏置为：", bias.eval())

    return None

if __name__ == "__main__":
    # myregression()
    restoremodel()

第一次训练后的数据 重新加载后的数据