[CI] Upgrade vllm to newest commit (#3182)

### What this PR does / why we need it?
Upgrade vLLM to newest commit

- Fix the aclgraph doesn't work problem, caused by
24fab45d96
- Fix PoolerOutput import error, caused by
755ed7b05b
- Fix the aclgraph weight load error to keep the same with torchair fix.
4492e3a554

### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
All test should pass


- vLLM version: v0.10.2
- vLLM main:
52d0cb8458

---------

Signed-off-by: wangxiyuan <wangxiyuan1007@gmail.com>

vllm_ascend/patch/worker/patch_common/patch_weight_loader.py

@@ -1,9 +1,6 @@
 import torch
 from torch.nn.parameter import Parameter
 from vllm.logger import init_logger
-# yapf: disable
-from vllm.model_executor.parameter import ModelWeightParameter
-# yapf: enable
 from vllm.model_executor.utils import set_weight_attrs
 from vllm.utils import GiB_bytes
 
@@ -16,27 +13,15 @@ def create_weights(self, layer: torch.nn.Module, input_size_per_partition: int,
                    output_partition_sizes: list[int], input_size: int,
                    output_size: int, params_dtype: torch.dtype,
                    **extra_weight_attrs):
-    from vllm_ascend.ascend_config import get_ascend_config
-    ascend_config = get_ascend_config()
     # This method creates unquantized linear weights.
     # The weights are not quantized, and they are not sharded.
     # The amount of memory allocated for the weights is
     # sum(output_partition_sizes) * input_size_per_partition.
     try:
-        if ascend_config.torchair_graph_config.enabled:
-            weight = Parameter(torch.empty(sum(output_partition_sizes),
-                                           input_size_per_partition,
-                                           dtype=params_dtype),
-                               requires_grad=False)
-        else:
-            weight_loader = extra_weight_attrs.pop("weight_loader")
-            weight = ModelWeightParameter(data=torch.empty(
-                sum(output_partition_sizes),
-                input_size_per_partition,
-                dtype=params_dtype),
-                                          input_dim=1,
-                                          output_dim=0,
-                                          weight_loader=weight_loader)
+        weight = Parameter(torch.empty(sum(output_partition_sizes),
+                                       input_size_per_partition,
+                                       dtype=params_dtype),
+                           requires_grad=False)
     except torch.cuda.OutOfMemoryError as e:
         logger.error("Failed to create unquantized linear weights: %s", e)
         if torch.cuda.is_available():
@@ -49,8 +34,7 @@ def create_weights(self, layer: torch.nn.Module, input_size_per_partition: int,
             "Failed to create unquantized linear weights. "
             "This may be caused by insufficient memory to allocate "
             "the weight.") from e
-    if ascend_config.torchair_graph_config.enabled:
-        set_weight_attrs(weight, {"input_dim": 1, "output_dim": 0})
+    set_weight_attrs(weight, {"input_dim": 1, "output_dim": 0})
     layer.register_parameter("weight", weight)
     set_weight_attrs(weight, extra_weight_attrs)