[BugFix][Core] Fix a bug running multi-modal with ascend_scheduler (#3675)

This PR fix the bug related with running multi-modal models with
AscendScheduler. This bug was introduced by PR #2372 by using the same
parameter names as vLLM with different default values. 

Currently I fix this bug by changing the default values of these two
parameters to align with vLLM. 

- vLLM version: v0.11.0rc3
- vLLM main:
17c540a993

Signed-off-by: hw_whx <wanghexiang7@huawei.com>
Co-authored-by: hw_whx <wanghexiang7@huawei.com>

vllm_ascend/core/schedule_config.py

@@ -26,7 +26,7 @@ MAX_INT = 2147483647
 @dataclass
 class AscendSchedulerConfig(SchedulerConfig):
     enable_chunked_prefill: bool = False
-    max_long_partial_prefills: int = MAX_INT
+    max_long_partial_prefills: int = 1
     long_prefill_token_threshold: int = MAX_INT
     policy: str = "fcfs"
     scheduler_cls: Union[str, Type[object]] = (
@@ -73,9 +73,9 @@ class AscendSchedulerConfig(SchedulerConfig):
                 "max_num_batched_tokens and makes vLLM reject longer "
                 "sequences. Please increase max_num_batched_tokens or "
                 "decrease max_model_len.")
-        # concurrent partial prefills. Default is inf
+        # concurrent partial prefills. Default is 1 meaning not enabled.
         if self.max_long_partial_prefills is None:
-            self.max_long_partial_prefills = MAX_INT
+            self.max_long_partial_prefills = 1
             self.long_prefill_token_threshold = MAX_INT
 
         if self.long_prefill_token_threshold is None or \
@@ -105,4 +105,4 @@ class AscendSchedulerConfig(SchedulerConfig):
         if getattr(self, "scheduler_delay_factor", 0) > 0:
             raise NotImplementedError(
                 "currently AscendScheduler doesn't support scheduler_delay_factor."
-            )
+            )