[Main2Main] Upgrade vllm commit to releases/v0.14.0 (#5911)

### What this PR does / why we need it?
Upgrade vllm commit to releases/v0.14.0

- Re-open cases in `tests/e2e/singlecard/pooling/test_scoring.py`, since
the errors before have been fixed by
https://github.com/vllm-project/vllm/pull/32243
### Does this PR introduce _any_ user-facing change?

### How was this patch tested?

- vLLM version: v0.13.0
- vLLM main:
11b6af5280

Signed-off-by: wjunLu <wjunlu217@gmail.com>

.github/workflows/_e2e_test.yaml

@@ -446,4 +446,4 @@ jobs:
           PYTORCH_NPU_ALLOC_CONF: max_split_size_mb:256
           VLLM_WORKER_MULTIPROC_METHOD: spawn
         run: |
-          pytest -sv --durations=0 tests/e2e/310p/test_offline_inference_parallel_310p.py
+          pytest -sv --durations=0 tests/e2e/310p/test_offline_inference_parallel_310p.py

.github/workflows/bot_pr_create.yaml

@@ -37,7 +37,7 @@ jobs:
     steps:
       - name: Get vLLM version
         run: |
-          VLLM_COMMIT=11b6af5280d6d6dfb8953af16e67b25f819b3be9
+          VLLM_COMMIT=2c24bc6996cb165fce92f780b388a5e39b3f4060
           echo "VLLM_COMMIT=https://github.com/vllm-project/vllm/commit/$VLLM_COMMIT" >> $GITHUB_ENV
 
       - name: Checkout repository

.github/workflows/pr_test_full.yaml

@@ -75,7 +75,7 @@ jobs:
     name: e2e-full
     strategy:
       matrix:
-        vllm_version: [11b6af5280d6d6dfb8953af16e67b25f819b3be9, v0.13.0]
+        vllm_version: [2c24bc6996cb165fce92f780b388a5e39b3f4060, v0.13.0]
     needs: [changes]
     if: ${{ needs.changes.outputs.e2e_tracker == 'true' }}
     uses: ./.github/workflows/_e2e_test.yaml

.github/workflows/pr_test_light.yaml

@@ -41,7 +41,7 @@ jobs:
   lint:
     uses: ./.github/workflows/_pre_commit.yml
     with:
-      vllm: 11b6af5280d6d6dfb8953af16e67b25f819b3be9
+      vllm: 2c24bc6996cb165fce92f780b388a5e39b3f4060
   changes:
     runs-on: linux-aarch64-a2-0
     outputs:
@@ -84,7 +84,7 @@ jobs:
     if: ${{ needs.lint.result == 'success' && (needs.changes.outputs.e2e_tracker == 'true' || needs.changes.outputs.ut_tracker == 'true') }}
     strategy:
       matrix:
-        vllm_version: [11b6af5280d6d6dfb8953af16e67b25f819b3be9, v0.13.0]
+        vllm_version: [2c24bc6996cb165fce92f780b388a5e39b3f4060, v0.13.0]
     uses: ./.github/workflows/_unit_test.yaml
     with:
       vllm: ${{ matrix.vllm_version }}
@@ -96,7 +96,7 @@ jobs:
     name: e2e-light
     strategy:
       matrix:
-        vllm_version: [11b6af5280d6d6dfb8953af16e67b25f819b3be9, v0.13.0]
+        vllm_version: [2c24bc6996cb165fce92f780b388a5e39b3f4060, v0.13.0]
     # 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/schedule_codecov_refresh.yaml

@@ -33,7 +33,7 @@ jobs:
     name: refresh codecov
     strategy:
       matrix:
-        vllm_version: [11b6af5280d6d6dfb8953af16e67b25f819b3be9]
+        vllm_version: [2c24bc6996cb165fce92f780b388a5e39b3f4060]
     uses: ./.github/workflows/_unit_test.yaml
     with:
       vllm: ${{ matrix.vllm_version }}

docs/source/community/versioning_policy.md

@@ -53,7 +53,7 @@ For main branch of vLLM Ascend, we usually make it compatible with the latest vL
 
 | vLLM Ascend | vLLM         | Python           | Stable CANN | PyTorch/torch_npu  |
 |-------------|--------------|------------------|-------------|--------------------|
-|     main    | 11b6af5280d6d6dfb8953af16e67b25f819b3be9, v0.13.0 tag | >= 3.10, < 3.12   | 8.3.RC2 | 2.8.0 / 2.8.0 |
+|     main    | 2c24bc6996cb165fce92f780b388a5e39b3f4060, v0.13.0 tag | >= 3.10, < 3.12   | 8.3.RC2 | 2.8.0 / 2.8.0 |
 
 ## Release cadence
 

tests/e2e/singlecard/pooling/test_scoring.py

@@ -5,8 +5,6 @@ import torch
 import torch.nn.functional as F
 from modelscope import snapshot_download  # type: ignore[import-untyped]
 
-from vllm_ascend.utils import vllm_version_is
-
 from tests.e2e.conftest import HfRunner, VllmRunner
 
 CROSS_ENCODER_MODELS = [
@@ -35,10 +33,6 @@ DTYPE = "half"
 def model_name(request):
     yield snapshot_download(request.param)
 
-@pytest.mark.skipif(
-    not vllm_version_is('0.13.0'),
-    reason="vLLM PR-32148 changed the behavior of cross scoring",
-)
 def test_cross_encoder_score_1_to_1(model_name):
     text_pair = [TEXTS_1[0], TEXTS_2[0]]
 
@@ -58,10 +52,6 @@ def test_cross_encoder_score_1_to_1(model_name):
     assert hf_outputs[0] == pytest.approx(vllm_outputs[0], rel=0.01)
 
 
-@pytest.mark.skipif(
-    not vllm_version_is('0.13.0'),
-    reason="vLLM PR-32148 changed the behavior of cross scoring",
-)
 def test_cross_encoder_score_1_to_N(model_name):
     text_pairs = [
         [TEXTS_1[0], TEXTS_2[0]],
@@ -85,10 +75,6 @@ def test_cross_encoder_score_1_to_N(model_name):
     assert hf_outputs[1] == pytest.approx(vllm_outputs[1], rel=0.01)
 
 
-@pytest.mark.skipif(
-    not vllm_version_is('0.13.0'),
-    reason="vLLM PR-32148 changed the behavior of cross scoring",
-)
 def test_cross_encoder_score_N_to_N(model_name):
     text_pairs = [
         [TEXTS_1[0], TEXTS_2[0]],