tensorflow 2.0学习笔记6

tensorflow学习笔记06

有错误还请各位指正！

自定义层

创建子类：tf.keras.Layer

两个方法：__init__:；call:前向传播算法

普通使用：

class Linear(tf.keras.layers.Layer):
    def __init__(self, units=32, input_dim=32):
        super(Linear, self).__init__()
        w_init = tf.random_normal_initializer()
        self.w = tf.Variable(
                 w_init(shape=(input_dim, units), dtype='float32'),
                 trainable=True
        )
        b_init = tf.zeros_initializer()
        self.b = tf.Variable(
                b_init(shape=(units, ), dtype='float32'),
                trainable=True
        )

    def call(self, inputs):
        return tf.matmul(inputs, self.w) + self.b

使用add_weight添加权重：

class Linear(tf.keras.layers.Layer):
    def __init__(self, units=32, input_dim=32):
        super(Linear, self).__init__()
        self.w = self.add_weight(
            shape=(input_dim, units),
            initializer='random_normal',
            trainable=True
        )
        self.b = self.add_weight(
            shape=(units, ),
            initializer='zeros',
            trainable=True
        )

    def call(self, inputs):
        return tf.matmul(inputs, self.w) + self.b

使用build添加权重：

不用固定输入的dims

class Linear(tf.keras.layers.Layer):
    def __init__(self, units=32):
        super(Linear, self).__init__()
        self.units = units
    
    def build(self, input_shape):
        self.w = self.add_weight(
            shape=(input_shape[-1], self.units),
            initializer='random_normal',
            trainable=True
        )
        self.b = self.add_weight(
            shape=(units, ),
            initializer='zeros',
            trainable=True
        )

    def call(self, inputs):
        return tf.matmul(inputs, self.w) + self.b

层的递归实现：

class MLPBlock(tf.keras.layers.Layer):
    def __init__(self):
        super(MLPBlock, self).__init__()
        self.lin_1 = Linear(32)
        self.lin_2 = Linear(64)
        self.lin_3 = Linear(1)

    def call(self, inputs):
        x = self.lin_1(inputs)
        x = tf.nn.relu(x)
        x = self.lin_2(x)
        x = tf.nn.relu(x)
        x = self.lin_3(x)

        return x

自定义模型

调用：tf.keras.Model

使用：

class MLP_model(tf.keras.Model):
    def __init__(self):
        super(MLP_model, self).__init__()
        self.lin_1 = Linear(32)
        self.lin_2 = Linear(64)
        self.lin_3 = Linear(1)

    def call(self, inputs):
        x = self.lin_1(inputs)
        x = tf.nn.relu(x)
        x = self.lin_2(x)
        x = tf.nn.relu(x)
        x = self.lin_3(x)

        return x

赏

谢谢你请我吃糖果

支付宝

微信

• 本文作者： Doted Wood
• 本文链接： http://example.com/2022/02/09/tensorflow/tensorflow-2-0学习笔记6/
• 版权声明： 版权归博主所有，转载请说明来源