在tensorflow 2.0.1中使用自定义层保存和加载模型

import tensorflow as tf

# 自定义层
class MyLayer(tf.keras.layers.Layer):
    def __init__(self, units):
        super(MyLayer, 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=(self.units,),
                                 initializer='zeros',
                                 trainable=True)

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

# 创建模型
model = tf.keras.Sequential([
    tf.keras.layers.Dense(64, activation='relu'),
    MyLayer(10),
    tf.keras.layers.Softmax()
])

# 保存模型
tf.keras.models.save_model(model, 'my_model')

import tensorflow as tf

# 加载模型
loaded_model = tf.keras.models.load_model('my_model')

# 使用加载的模型进行预测
predictions = loaded_model.predict(x_test)