This change separates the general DP padding logic from the existing implementation specific to TorchAir. Make sure server do not hang when batch size < DP size.

Signed-off-by: Yizhou Liu <[email protected]>

Author: yiz-liu

vllm_ascend/worker/model_runner_v1.py

@@ -1930,6 +1930,11 @@ def _dummy_run(
             )
 
         # Padding for DP
+        num_pad, num_tokens_across_dp_native = self.get_dp_padding(
+            num_tokens)
+        # num_tokens += num_pad  ## Uncomment this after TorchAir is removed
+
+        # Padding for DP (for TorchAir)
         (num_tokens, num_tokens_across_dp, with_prefill,
          _) = self._get_forward_metadata_across_dp_and_pad(
              num_tokens, with_prefill, False)