补充mrope创建部分

vllm/v1/worker/gpu_model_runner.py

@@ -611,9 +611,15 @@ class GPUModelRunner(
             # identical position IDs, making M-RoPE functionally equivalent to
             # 1D-RoPE.
             # See page 5 of https://arxiv.org/abs/2409.12191
-            self.mrope_positions = self._make_buffer(
-                (3, self.max_num_tokens + 1), dtype=torch.int64
-            )
+            if self.use_1d_mrope:
+                self.mrope_positions = self._make_buffer(
+                    3 * (self.max_num_tokens + 1), dtype=torch.int64
+                )
+            else:
+                self.mrope_positions = self._make_buffer(
+                    (3, self.max_num_tokens + 1), dtype=torch.int64
+                )
+
 
         # Only relevant for models using XD-RoPE (e.g, HunYuan-VL)
         if self.uses_xdrope_dim > 0: