Update Optional[x] -> x | None and Union[x, y] to x | y (#26633)

Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>

vllm/config/parallel.py

@@ -3,7 +3,7 @@
 
 import hashlib
 import os
-from typing import TYPE_CHECKING, Any, Literal, Optional, Union
+from typing import TYPE_CHECKING, Any, Literal
 
 import torch
 from pydantic import Field, model_validator
@@ -75,7 +75,7 @@ class ParallelConfig:
     """Number of local data parallel groups."""
     data_parallel_rank: int = 0
     """Rank of the data parallel group."""
-    data_parallel_rank_local: Optional[int] = None
+    data_parallel_rank_local: int | None = None
     """Local rank of the data parallel group,
     set only in SPMD mode."""
     data_parallel_master_ip: str = "127.0.0.1"
@@ -113,24 +113,24 @@ class ParallelConfig:
       with 4 experts and 2 ranks, rank 0 will have experts [0, 2] and rank 1
       will have experts [1, 3]. This strategy can help improve load balancing
       for grouped expert models with no redundant experts."""
-    num_redundant_experts: Optional[int] = None
+    num_redundant_experts: int | None = None
     """`num_redundant_experts` is deprecated and has been replaced with
     `eplb_config.num_redundant_experts`. This will be removed in v0.12.0.
     Please use `eplb_config.num_redundant_experts` instead."""
-    eplb_window_size: Optional[int] = None
+    eplb_window_size: int | None = None
     """`eplb_window_size` is deprecated and has been replaced with
     `eplb_config.window_size`. This will be removed in v0.12.0.
     Please use `eplb_config.window_size` instead."""
-    eplb_step_interval: Optional[int] = None
+    eplb_step_interval: int | None = None
     """`eplb_step_interval` is deprecated and has been replaced with
     `eplb_config.step_interval`. This will be removed in v0.12.0.
     Please use `eplb_config.step_interval` instead."""
-    eplb_log_balancedness: Optional[bool] = None
+    eplb_log_balancedness: bool | None = None
     """`eplb_log_balancedness` is deprecated and has been replaced with
     `eplb_config.log_balancedness`. This will be removed in v0.12.0.
     Please use `eplb_config.log_balancedness` instead."""
 
-    max_parallel_loading_workers: Optional[int] = None
+    max_parallel_loading_workers: int | None = None
     """Maximum number of parallel loading workers when loading model
     sequentially in multiple batches. To avoid RAM OOM when using tensor
     parallel and large models."""
@@ -159,15 +159,15 @@ class ParallelConfig:
     ray_workers_use_nsight: bool = False
     """Whether to profile Ray workers with nsight, see https://docs.ray.io/en/latest/ray-observability/user-guides/profiling.html#profiling-nsight-profiler."""
 
-    ray_runtime_env: Optional[RuntimeEnv] = None
+    ray_runtime_env: RuntimeEnv | None = None
     """Ray runtime environment to pass to distributed workers."""
 
-    placement_group: Optional[PlacementGroup] = None
+    placement_group: PlacementGroup | None = None
     """ray distributed model workers placement group."""
 
-    distributed_executor_backend: Optional[
-        Union[str, DistributedExecutorBackend, type[ExecutorBase]]
-    ] = None
+    distributed_executor_backend: (
+        str | DistributedExecutorBackend | type[ExecutorBase] | None
+    ) = None
     """Backend to use for distributed model
     workers, either "ray" or "mp" (multiprocessing). If the product
     of pipeline_parallel_size and tensor_parallel_size is less than
@@ -306,7 +306,7 @@ class ParallelConfig:
         )
 
         max_retries = 5
-        last_exc: Optional[Exception] = None
+        last_exc: Exception | None = None
         for _ in range(max_retries):
             try:
                 # use gloo since the engine process might not have cuda device