[Bugfix] Disable multi-modal preprocessor cache for DP (#21896)

Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
2025-08-01 15:03:56 +08:00
parent 79731a79f0
commit b4e081cb15
3 changed files with 21 additions and 2 deletions
--- a/vllm/entrypoints/cli/serve.py
+++ b/vllm/entrypoints/cli/serve.py
@@ -167,8 +167,9 @@ def run_multi_api_server(args: argparse.Namespace):

        if model_config.is_multimodal_model and not (
                orig_disable_mm_preprocessor_cache):
-            logger.warning("Multi-model preprocessor cache will be disabled "
-                           "for api_server_count > 1")
+            logger.warning(
+                "Multi-modal preprocessor cache is not compatible "
+                "with api_server_count > 1, so the cache will be disabled.")

    executor_class = Executor.get_class(vllm_config)
    log_stats = not engine_args.disable_log_stats