符号 self(x) 的作用是什么？

Question

我正在从 'Deep Reinforcement Learning Hands On' code 学习分布式强化学习。模型class中有一个方法：

    def both(self, x):
        cat_out = self(x)
        probs = self.apply_softmax(cat_out)
        weights = probs * self.supports
        res = weights.sum(dim=2)
        return cat_out, res

那个 self(x) do/mean 是什么？

Answer 1

它将在实例上调用 __call__ 方法。请参阅此演示：

class A:
    def __call__(self, x):
        print("called instance")
        return x + 3

    def both(self, x):
        val = self(x)
        print("value:", val)


a = A()
a.both(5)
print()
a(123)
a.__call__(123)

Output:

called instance
value: 8

called instance
called instance

val = self(x)与val = self.__call__(x)相同

符号 self(x) 的作用是什么？

What does the notation self(x) do?

python

self

reinforcement-learning

deep-learning