[Deprecation] Deprecate seed=None (#29185)

Signed-off-by: DarkLight1337 <[email protected]>

Author: DarkLight1337

vllm/engine/arg_utils.py

@@ -367,7 +367,7 @@ class EngineArgs:
     config_format: str = ModelConfig.config_format
     dtype: ModelDType = ModelConfig.dtype
     kv_cache_dtype: CacheDType = CacheConfig.cache_dtype
-    seed: int | None = None
+    seed: int | None = 0
     max_model_len: int | None = ModelConfig.max_model_len
     cuda_graph_sizes: list[int] | None = CompilationConfig.cudagraph_capture_sizes
     cudagraph_capture_sizes: list[int] | None = (
@@ -1192,6 +1192,12 @@ def create_model_config(self) -> ModelConfig:
         # VLLM_ENABLE_V1_MULTIPROCESSING=0), so setting a seed here
         # doesn't affect the user process.
         if self.seed is None:
+            logger.warning_once(
+                "`seed=None` is equivalent to `seed=0` in V1 Engine. "
+                "You will no longer be allowed to pass `None` in v0.13.",
+                scope="local",
+            )
+
             self.seed = 0
             if not envs.VLLM_ENABLE_V1_MULTIPROCESSING:
                 logger.warning(
@@ -1203,28 +1209,31 @@ def create_model_config(self) -> ModelConfig:
                 )
 
         if self.disable_mm_preprocessor_cache:
-            logger.warning(
+            logger.warning_once(
                 "`--disable-mm-preprocessor-cache` is deprecated "
                 "and will be removed in v0.13. "
                 "Please use `--mm-processor-cache-gb 0` instead.",
+                scope="local",
             )
 
             self.mm_processor_cache_gb = 0
         elif envs.VLLM_MM_INPUT_CACHE_GIB != 4:
-            logger.warning(
+            logger.warning_once(
                 "VLLM_MM_INPUT_CACHE_GIB` is deprecated "
                 "and will be removed in v0.13. "
                 "Please use `--mm-processor-cache-gb %d` instead.",
                 envs.VLLM_MM_INPUT_CACHE_GIB,
+                scope="local",
             )
 
             self.mm_processor_cache_gb = envs.VLLM_MM_INPUT_CACHE_GIB
 
         if self.enable_multimodal_encoder_data_parallel:
-            logger.warning(
+            logger.warning_once(
                 "--enable-multimodal-encoder-data-parallel` is deprecated "
                 "and will be removed in v0.13. "
-                "Please use `--mm-encoder-tp-mode data` instead."
+                "Please use `--mm-encoder-tp-mode data` instead.",
+                scope="local",
             )
 
             self.mm_encoder_tp_mode = "data"