Clean up comment

Author: yanxi-chen

trinity/trainer/verl/dp_actor.py

@@ -118,7 +118,7 @@ def update_policy(self, data: DataProto):  # noqa: C901
                     # calculate the total number of response tokens in the minibatch
                     mini_batch_token_num = torch.sum(
                         mini_batch.batch["response_mask"].to(get_device_id())
-                    ).item()  # TODO: double check this calculation
+                    ).item()
 
                 self.actor_optimizer.zero_grad()