3cbccdbab2c558642cb11e45eaf7f568968d688a,slm_lab/agent/algorithm/actor_critic.py,ActorCritic,update,#ActorCritic#,355

Before Change

    @lab_api
    def update(self):
        for net_name in self.net_names:
            net = getattr(self, net_name)
            self.body.grad_norms.extend(net.grad_norms)
        self.body.explore_var = self.explore_var_scheduler.update(self, self.body.env.clock)
        if self.self.entropy_coef_spec is not None:
            self.body.entropy_coef = self.entropy_coef_scheduler.update(self, self.body.env.clock)
        return self.body.explore_var

After Change

    @lab_api
    def update(self):
        net_util.try_store_grad_norm(self)
        self.body.explore_var = self.explore_var_scheduler.update(self, self.body.env.clock)
        if self.self.entropy_coef_spec is not None:
            self.body.entropy_coef = self.entropy_coef_scheduler.update(self, self.body.env.clock)
        return self.body.explore_var

Instances