import tensorflow as tf # 设置输入层节点数、隐层节点数 in_nodes=784 h1_nodes=100 h2_nodes=100 h3_nodes=50 # 定义输入、输出、prob的placeholder x=tf.keras.Input(shape=(in_nodes,)) y_=tf.keras.Input(shape=(10,)) prob=tf.keras.Input(shape=()) # 设置第一隐层 w1=weight([in_nodes, h1_nodes], 0.1) b1=tf.Variable(tf.zeros([h1_nodes])) h1=tf.nn.relu(tf.matmul(x,w1)+b1) # 设置第二隐层 w2=weight([h1_nodes, h2_nodes], 0.0) b2=tf.Variable(tf.zeros([h2_nodes])) h2=tf.nn.relu(tf.matmul(h1,w2)+b2) h2_drop=tf.nn.dropout(h2, rate=prob) # 设置第三隐层 w3=weight([h2_nodes, h3_nodes], 0.0) b3=tf.Variable(tf.zeros([h3_nodes])) h3=tf.nn.relu(tf.matmul(h2_drop,w3)+b3) h3_drop=tf.nn.dropout(h3, rate=prob) # 设置softmax输出层 w4=weight([h3_nodes, 10], 0.0) b4=tf.Variable(tf.zeros([10])) y=tf.nn.softmax(tf.matmul(h3_drop,w4)+b4)改为可在tensorflow2.x使用

您好，将上述代码改写成 TensorFlow 2.x 版本的代码如下：

import tensorflow as tf

class MyModel(tf.keras.Model):
    def __init__(self, in_nodes=784, h1_nodes=100, h2_nodes=100, h3_nodes=50):
        super(MyModel, self).__init__()
        
        self.w1 = tf.Variable(tf.random.truncated_normal([in_nodes, h1_nodes], stddev=0.1))
        self.b1 = tf.Variable(tf.zeros([h1_nodes]))
        self.w2 = tf.Variable(tf.random.truncated_normal([h1_nodes, h2_nodes], stddev=0.1))
        self.b2 = tf.Variable(tf.zeros([h2_nodes]))
        self.w3 = tf.Variable(tf.random.truncated_normal([h2_nodes, h3_nodes], stddev=0.1))
        self.b3 = tf.Variable(tf.zeros([h3_nodes]))
        self.w4 = tf.Variable(tf.random.truncated_normal([h3_nodes, 10], stddev=0.1))
        self.b4 = tf.Variable(tf.zeros([10]))
        
    def call(self, inputs, prob):
        x = inputs
        y_ = tf.cast(inputs, tf.float32)
        h1 = tf.nn.relu(tf.matmul(x, self.w1) + self.b1)
        h2 = tf.nn.relu(tf.matmul(h1, self.w2) + self.b2)
        h2_drop = tf.nn.dropout(h2, rate=prob)
        h3 = tf.nn.relu(tf.matmul(h2_drop, self.w3) + self.b3)
        h3_drop = tf.nn.dropout(h3, rate=prob)
        y = tf.nn.softmax(tf.matmul(h3_drop, self.w4) + self.b4)
        return y

model = MyModel()
x = tf.keras.Input(shape=(784,))
prob = tf.keras.Input(shape=())
y = model(x, prob)

注意到上述代码中使用了 `tf.keras.Model` 类来定义模型，`__init__` 方法中定义了模型的所有变量，`call` 方法中定义了前向传播过程。同时，使用 `tf.keras.Input` 定义了输入数据的形状。