add comment

Signed-off-by: Leo Tian <leo.tian@centml.ai>

Author: leo-cf-tian

vllm/v1/worker/gpu_input_batch.py

@@ -51,6 +51,7 @@ def get_token_id(self, idx: int) -> int:
             return self.prompt_token_ids[idx]
         elif idx - self.num_prompt_tokens < len(self.output_token_ids):
             return self.output_token_ids[idx - self.num_prompt_tokens]
+        # This is now precomputed, so we create a fallback if the idx is invalid
         else:
             return -1  # Invalid token id