1.类
- 属性、方法、构造函数
2.类的继承
- 基类、派生类
- super()方法:子类对象调用父类已被覆盖的方法
- call()方法:实例名()将调用call()方法
3.模块、层和模型
-
层的设计:
class Dense(tf.Module): def __init__(self, in_features, out_features, name=none): super().__init__(name=name) self.w = tf.Variable( tf.random.normal([in_features, out_features]), name='w') self.b = tf.Variable(tf.zeros([out_features]), name='b') def __call__(self, x): y = tf.matmul(x, self.w) + self.b return tf.nn.relu(y)
-
从层到模型:Sequential模型
-
动态决定张量维度:
is_built
-
模型存储和恢复:
new_checkpoint = tf.train.Checkpoint(model=new_model) new_checkpoint.restore("my_checkpoint")
-
装饰器:
@tf.function
用作call函数的规范调用
4.tensorboard
可视化
5.模型的存储
tf.saved_model.save(my_model, 'path')
tf.saved_model.load
6.计算图
7.训练流程