add torch.no_grad decorators

src/lerobot/policies/multi_task_dit/modeling_multi_task_dit.py

@@ -760,6 +760,7 @@ class MultiTaskDiTPolicy(PreTrainedPolicy):
 
         self._queues["task"] = deque(maxlen=self.config.n_obs_steps)
 
+    @torch.no_grad()
     def predict_action_chunk(self, batch: dict[str, Tensor]) -> Tensor:
         """Predict a chunk of actions given environment observations"""
         self.eval()
@@ -778,6 +779,7 @@ class MultiTaskDiTPolicy(PreTrainedPolicy):
         actions = self._generate_actions(batch)
         return actions
 
+    @torch.no_grad()
     def select_action(self, batch: dict[str, Tensor]) -> Tensor:
         """Select a single action given environment observations"""
         if ACTION in batch: