[Bugfix] Fix wrong CLI defaults for dynamic `SchedulerConfig` fields (#28872)

Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>

tests/entrypoints/openai/test_enable_force_include_usage.py

@@ -17,7 +17,7 @@ def chat_server_with_force_include_usage(request):  # noqa: F811
         "128",
         "--enforce-eager",
         "--max-num-seqs",
-        "1",
+        "4",
         "--enable-force-include-usage",
         "--port",
         "55857",
@@ -78,7 +78,7 @@ def transcription_server_with_force_include_usage():
         "--dtype",
         "bfloat16",
         "--max-num-seqs",
-        "1",
+        "4",
         "--enforce-eager",
         "--enable-force-include-usage",
         "--gpu-memory-utilization",

vllm/engine/arg_utils.py

@@ -1046,10 +1046,18 @@ class EngineArgs:
             description=SchedulerConfig.__doc__,
         )
         scheduler_group.add_argument(
-            "--max-num-batched-tokens", **scheduler_kwargs["max_num_batched_tokens"]
+            "--max-num-batched-tokens",
+            **{
+                **scheduler_kwargs["max_num_batched_tokens"],
+                "default": None,
+            },
         )
         scheduler_group.add_argument(
-            "--max-num-seqs", **scheduler_kwargs["max_num_seqs"]
+            "--max-num-seqs",
+            **{
+                **scheduler_kwargs["max_num_seqs"],
+                "default": None,
+            },
         )
         scheduler_group.add_argument(
             "--max-num-partial-prefills", **scheduler_kwargs["max_num_partial_prefills"]
@@ -1071,7 +1079,11 @@ class EngineArgs:
             "--scheduling-policy", **scheduler_kwargs["policy"]
         )
         scheduler_group.add_argument(
-            "--enable-chunked-prefill", **scheduler_kwargs["enable_chunked_prefill"]
+            "--enable-chunked-prefill",
+            **{
+                **scheduler_kwargs["enable_chunked_prefill"],
+                "default": None,
+            },
         )
         scheduler_group.add_argument(
             "--disable-chunked-mm-input", **scheduler_kwargs["disable_chunked_mm_input"]