[CI] Mock spawn for vlm tests (#5279)

### What this PR does / why we need it?
Using `spawn` in continuous testing scenarios
### Does this PR introduce _any_ user-facing change?

### How was this patch tested?

- vLLM version: release/v0.13.0
- vLLM main:
ad32e3e19c

---------

Signed-off-by: wangli <wangli858794774@gmail.com>

.github/workflows/_e2e_test.yaml

@@ -86,6 +86,7 @@ jobs:
       - name: Run vllm-project/vllm-ascend test
         env:
           PYTORCH_NPU_ALLOC_CONF: max_split_size_mb:256
+          VLLM_WORKER_MULTIPROC_METHOD: spawn
         if: ${{ inputs.type == 'light' }}
         run: |
           # pytest -sv --durations=0 tests/e2e/singlecard/test_aclgraph_accuracy.py
@@ -177,6 +178,8 @@ jobs:
 
       - name: Run vllm-project/vllm-ascend test (non triton)
         if: ${{ inputs.type == 'full' }}
+        env:
+          VLLM_WORKER_MULTIPROC_METHOD: spawn
         run: |
           pytest -sv --durations=0 tests/e2e/multicard/test_aclgraph_capture_replay.py
 

tests/e2e/singlecard/test_vlm.py

@@ -20,6 +20,9 @@
 
 Run `pytest tests/test_offline_inference.py`.
 """
+import os
+from unittest.mock import patch
+
 from vllm import SamplingParams
 from vllm.assets.audio import AudioAsset
 from vllm.assets.image import ImageAsset
@@ -27,6 +30,7 @@ from vllm.assets.image import ImageAsset
 from tests.e2e.conftest import VllmRunner
 
 
+@patch.dict(os.environ, {"VLLM_WORKER_MULTIPROC_METHOD": "spawn"})
 def test_multimodal_vl(vl_config):
     image = ImageAsset("cherry_blossom").pil_image.convert("RGB")
 
@@ -57,6 +61,7 @@ def test_multimodal_vl(vl_config):
             assert output_str, "Generated output should not be empty."
 
 
+@patch.dict(os.environ, {"VLLM_WORKER_MULTIPROC_METHOD": "spawn"})
 def test_multimodal_audio():
     audio_prompt = "".join([
         f"Audio {idx+1}: <|audio_bos|><|AUDIO|><|audio_eos|>\n"