Method update_distribution

humanoidverse/agents/modules/ppo_modules.py:63–65

(self, actor_obs)

        return self.distribution.entropy().sum(dim=-1)

    def update_distribution(self, actor_obs):
        mean = self.actor(actor_obs)
        self.distribution = Normal(mean, mean*0. + self.std)

    def act(self, actor_obs, **kwargs):
        self.update_distribution(actor_obs)