[CI] Upgrade vllm version - 0925 (#3167)

Upgrade vLLM to newest commit.

1. Remove the useless func get_state_cls, it has been removed from vLLM
already.
e6750d0b18
2. Fix ut broken by
6160ba4151


- vLLM version: v0.10.2
- vLLM main:
b1068903fd

---------

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

.github/workflows/format_pr_body.yaml

@@ -36,7 +36,7 @@ jobs:
 
       - name: Get vLLM version
         run: |
-          VLLM_COMMIT=b1068903fdca26cf6b4a1a51a32c3365ce3ac636
+          VLLM_COMMIT=52d0cb845866869d587fc013a7c59e60a86ebcf2
           echo "VLLM_COMMIT=https://github.com/vllm-project/vllm/commit/$VLLM_COMMIT" >> $GITHUB_ENV
 
       - name: Checkout repository

.github/workflows/vllm_ascend_test.yaml

@@ -42,7 +42,7 @@ jobs:
   lint:
     uses: ./.github/workflows/pre-commit.yml
     with:
-      vllm: b1068903fdca26cf6b4a1a51a32c3365ce3ac636
+      vllm: 52d0cb845866869d587fc013a7c59e60a86ebcf2
 
   changes:
     runs-on: ubuntu-latest
@@ -83,7 +83,7 @@ jobs:
         VLLM_USE_MODELSCOPE: True
     strategy:
       matrix:
-        vllm_version: [b1068903fdca26cf6b4a1a51a32c3365ce3ac636, v0.10.2]
+        vllm_version: [52d0cb845866869d587fc013a7c59e60a86ebcf2, v0.10.2]
     steps:
       - name: Install packages
         run: |
@@ -138,7 +138,7 @@ jobs:
     name: e2e-light
     strategy:
       matrix:
-        vllm_version: [b1068903fdca26cf6b4a1a51a32c3365ce3ac636, v0.10.2]
+        vllm_version: [52d0cb845866869d587fc013a7c59e60a86ebcf2, v0.10.2]
     # Note (yikun): If CI resource are limited we can split job into two chain jobs
     needs: [lint, changes]
     # only trigger e2e test after lint passed and the change is e2e related with pull request.

.github/workflows/vllm_ascend_test_full.yaml

@@ -68,7 +68,7 @@ jobs:
     name: e2e-full
     strategy:
       matrix:
-        vllm_version: [b1068903fdca26cf6b4a1a51a32c3365ce3ac636, v0.10.2]
+        vllm_version: [52d0cb845866869d587fc013a7c59e60a86ebcf2, v0.10.2]
     needs: [changes]
     if: ${{ needs.changes.outputs.e2e_tracker == 'true' }}
     uses: ./.github/workflows/_e2e_test.yaml

tests/ut/attention/test_attention_v1.py

@@ -7,8 +7,7 @@ from vllm_ascend.attention.attention_v1 import (AscendAttentionBackend,
                                                 AscendAttentionBackendImpl,
                                                 AscendAttentionMetadataBuilder,
                                                 AscendAttentionState,
-                                                AscendMetadata,
-                                                CommonAttentionState)
+                                                AscendMetadata)
 from vllm_ascend.attention.utils import AscendCommonAttentionMetadata
 
 
@@ -25,10 +24,6 @@ class TestAscendAttentionBackend(TestBase):
         self.assertEqual(AscendAttentionBackend.get_metadata_cls(),
                          AscendMetadata)
 
-    def test_get_state_cls(self):
-        self.assertEqual(AscendAttentionBackend.get_state_cls(),
-                         CommonAttentionState)
-
     def test_get_builder_cls(self):
         self.assertEqual(AscendAttentionBackend.get_builder_cls(),
                          AscendAttentionMetadataBuilder)

tests/ut/ops/test_fused_ops.py

@@ -201,6 +201,8 @@ def default_moe_config():
 def moe_method(mock_dist_env):
     moe = MagicMock()
     moe.moe_parallel_config.return_value = MagicMock(ep_size=4)
+    moe.moe_parallel_config.use_ep = False
+    moe.moe_parallel_config.dp_size = 1
     return AscendUnquantizedFusedMoEMethod(moe)
 
 

tests/ut/torchair/ops/test_torchair_fused_moe.py

@@ -153,6 +153,8 @@ def default_moe_config():
 def moe_method(mock_dist_env):
     moe = MagicMock()
     moe.moe_parallel_config.return_value = MagicMock(ep_size=4)
+    moe.moe_parallel_config.use_ep = False
+    moe.moe_parallel_config.dp_size = 1
     return TorchairAscendUnquantizedFusedMoEMethod(moe)
 
 

vllm_ascend/attention/attention_v1.py

@@ -24,7 +24,6 @@ import torch.nn as nn
 import torch_npu
 from vllm.attention.backends.abstract import (AttentionBackend, AttentionImpl,
                                               AttentionLayer, AttentionType)
-from vllm.attention.backends.utils import CommonAttentionState
 from vllm.config import VllmConfig
 from vllm.forward_context import ForwardContext, get_forward_context
 from vllm.utils import cdiv, direct_register_custom_op
@@ -56,10 +55,6 @@ class AscendAttentionBackend(AttentionBackend):
     def get_metadata_cls() -> Type["AscendMetadata"]:
         return AscendMetadata
 
-    @staticmethod
-    def get_state_cls() -> Type["CommonAttentionState"]:
-        return CommonAttentionState
-
     @staticmethod
     def get_builder_cls() -> type["AscendAttentionMetadataBuilder"]:
         return AscendAttentionMetadataBuilder