fix(rl): improve action processing for discrete and continuous actions

(cherry picked from commit f887ab3f6a)

src/lerobot/rl/actor.py

@@ -291,7 +291,19 @@ def act_with_policy(
         with policy_timer:
             normalized_observation = preprocessor.process_observation(observation)
             action = policy.select_action(batch=normalized_observation)
-            action = postprocessor.process_action(action)
+            # Unnormalize only the continuous part. When `num_discrete_actions` is set,
+            # `select_action` concatenates an argmax index in env space at the last dim;
+            # action stats cover the continuous dims only, so feeding the full vector to
+            # the unnormalizer would shape-mismatch and would also corrupt the discrete
+            # index by treating it as a normalized value.
+            if cfg.policy.num_discrete_actions is not None:
+                continuous_action = postprocessor.process_action(action[..., :-1])
+                discrete_action = action[..., -1:].to(
+                    device=continuous_action.device, dtype=continuous_action.dtype
+                )
+                action = torch.cat([continuous_action, discrete_action], dim=-1)
+            else:
+                action = postprocessor.process_action(action)
         policy_fps = policy_timer.fps_last
 
         log_policy_frequency_issue(policy_fps=policy_fps, cfg=cfg, interaction_step=interaction_step)