Fix critic_observations into privileged_observations in transition class (#6)

Author: GiulioRomualdi

amp_rsl_rl/algorithms/amp_ppo.py

@@ -224,7 +224,7 @@ def act(self, obs: torch.Tensor, critic_obs: torch.Tensor) -> torch.Tensor:
         self.transition.action_sigma = self.actor_critic.action_std.detach()
         # Record the observations before taking an environment step.
         self.transition.observations = obs
-        self.transition.critic_observations = critic_obs
+        self.transition.privileged_observations = critic_obs
         return self.transition.actions
 
     def act_amp(self, amp_obs: torch.Tensor) -> None: