【tensorflow2.0】低阶api--张量操作、计算图、自动微分

下面的范例使用TensorFlow的低阶API实现线性回归模型。

低阶API主要包括张量操作，计算图和自动微分。

import tensorflow as tf
 
# 打印时间分割线
@tf.function
def printbar():
    ts = tf.timestamp()
    today_ts = ts%(24*60*60)
 
    hour = tf.cast(today_ts//3600+8,tf.int32)%tf.constant(24)
    minite = tf.cast((today_ts%3600)//60,tf.int32)
    second = tf.cast(tf.floor(today_ts%60),tf.int32)
 
    def timeformat(m):
        if tf.strings.length(tf.strings.format("{}",m))==1:
            return(tf.strings.format("0{}",m))
        else:
            return(tf.strings.format("{}",m))
 
    timestring = tf.strings.join([timeformat(hour),timeformat(minite),
                timeformat(second)],separator = ":")
    tf.print("=========="*8,end = "")
    tf.print(timestring)
 
# 样本数量
n = 400
 
# 生成测试用数据集
X = tf.random.uniform([n,2],minval=-10,maxval=10) 
w0 = tf.constant([[2.0],[-1.0]])
b0 = tf.constant(3.0)
Y = X@w0 + b0 + tf.random.normal([n,1],mean = 0.0,stddev= 2.0)  # @表示矩阵乘法,增加正态扰动
 
# 使用动态图调试
 
w = tf.Variable(tf.random.normal(w0.shape))
b = tf.Variable(0.0)
 
def train(epoches):
    for epoch in tf.range(1,epoches+1):
        with tf.GradientTape() as tape:
            #正向传播求损失
            Y_hat = X@w + b
            loss = tf.squeeze(tf.transpose(Y-Y_hat)@(Y-Y_hat))/(2.0*n)   
 
        # 反向传播求梯度
        dloss_dw,dloss_db = tape.gradient(loss,[w,b])
        # 梯度下降法更新参数
        w.assign(w - 0.001*dloss_dw)
        b.assign(b - 0.001*dloss_db)
        if epoch%1000 == 0:
            printbar()
            tf.print("epoch =",epoch," loss =",loss,)
            tf.print("w =",w)
            tf.print("b =",b)
            tf.print("")
 
train(5000)

结果：

================================================================================15:18:17
epoch = 1000  loss = 2.66289544
w = [[2.0176034]
 [-1.02091444]]
b = 1.92718041

================================================================================15:18:19
epoch = 2000  loss = 2.12707591
w = [[2.01378]
 [-1.01979101]]
b = 2.63039422

================================================================================15:18:21
epoch = 3000  loss = 2.05447602
w = [[2.01237178]
 [-1.01937926]]
b = 2.88924217

================================================================================15:18:23
epoch = 4000  loss = 2.04463911
w = [[2.01185489]
 [-1.01922464]]
b = 2.98452425

================================================================================15:18:24
epoch = 5000  loss = 2.04330635
w = [[2.01166272]
 [-1.01917028]]
b = 3.01959634

转换成静态图加速：

## 使用autograph机制转换成静态图加速
 
w = tf.Variable(tf.random.normal(w0.shape))
b = tf.Variable(0.0)
 
@tf.function
def train(epoches):
    for epoch in tf.range(1,epoches+1):
        with tf.GradientTape() as tape:
            #正向传播求损失
            Y_hat = X@w + b
            loss = tf.squeeze(tf.transpose(Y-Y_hat)@(Y-Y_hat))/(2.0*n)   
 
        # 反向传播求梯度
        dloss_dw,dloss_db = tape.gradient(loss,[w,b])
        # 梯度下降法更新参数
        w.assign(w - 0.001*dloss_dw)
        b.assign(b - 0.001*dloss_db)
        if epoch%1000 == 0:
            printbar()
            tf.print("epoch =",epoch," loss =",loss,)
            tf.print("w =",w)
            tf.print("b =",b)
            tf.print("")
train(5000)

结果：

================================================================================15:19:50
epoch = 1000  loss = 2.6668539
w = [[2.01762223]
 [-1.02092016]]
b = 1.92363214

================================================================================15:19:51
epoch = 2000  loss = 2.12761354
w = [[2.01378703]
 [-1.01979291]]
b = 2.6290853

================================================================================15:19:52
epoch = 3000  loss = 2.0545485
w = [[2.0123744]
 [-1.01938]]
b = 2.888762

================================================================================15:19:53
epoch = 4000  loss = 2.04464912
w = [[2.01185584]
 [-1.019225]]
b = 2.98434567

================================================================================15:19:54
epoch = 5000  loss = 2.04330778
w = [[2.0116632]
 [-1.0191704]]
b = 3.01952934

参考：

开源电子书地址：https://lyhue1991.github.io/eat_tensorflow2_in_30_days/

GitHub 项目地址：https://github.com/lyhue1991/eat_tensorflow2_in_30_days