Merge tag 'v0.18.1rc0' into v0.18.1rc0-ori

tests/entrypoints/instrumentator/test_basic.py → tests/entrypoints/serve/instrumentator/test_basic.py

@@ -11,11 +11,10 @@ import pytest_asyncio
 import requests
 from fastapi import Request
 
+from tests.utils import RemoteOpenAIServer
 from vllm.v1.engine.exceptions import EngineDeadError
 from vllm.version import __version__ as VLLM_VERSION
 
-from ...utils import RemoteOpenAIServer
-
 MODEL_NAME = "Qwen/Qwen3-0.6B"
 
 
@@ -28,7 +27,7 @@ def server_args(request: pytest.FixtureRequest) -> list[str]:
     >>> @pytest.mark.parametrize(
     >>>     "server_args",
     >>>     [
-    >>>         ["--disable-frontend-multiprocessing"],
+    >>>         ["--max-model-len", "10100"],
     >>>         [
     >>>             "--model=NousResearch/Hermes-3-Llama-3.1-70B",
     >>>             "--enable-auto-tool-choice",
@@ -40,7 +39,7 @@ def server_args(request: pytest.FixtureRequest) -> list[str]:
     >>>     ...
 
     This will run `test_foo` twice with servers with:
-    - `--disable-frontend-multiprocessing`
+    - `--max-model-len 10100`
     - `--model=NousResearch/Hermes-3-Llama-3.1-70B --enable-auto-tool-choice`.
 
     """
@@ -79,17 +78,6 @@ async def client(server):
         yield async_client
 
 
-@pytest.mark.parametrize(
-    "server_args",
-    [
-        pytest.param([], id="default-frontend-multiprocessing"),
-        pytest.param(
-            ["--disable-frontend-multiprocessing"],
-            id="disable-frontend-multiprocessing",
-        ),
-    ],
-    indirect=True,
-)
 @pytest.mark.asyncio
 async def test_show_version(server: RemoteOpenAIServer):
     response = requests.get(server.url_for("version"))
@@ -98,17 +86,6 @@ async def test_show_version(server: RemoteOpenAIServer):
     assert response.json() == {"version": VLLM_VERSION}
 
 
-@pytest.mark.parametrize(
-    "server_args",
-    [
-        pytest.param([], id="default-frontend-multiprocessing"),
-        pytest.param(
-            ["--disable-frontend-multiprocessing"],
-            id="disable-frontend-multiprocessing",
-        ),
-    ],
-    indirect=True,
-)
 @pytest.mark.asyncio
 async def test_check_health(server: RemoteOpenAIServer):
     response = requests.get(server.url_for("health"))
@@ -119,13 +96,7 @@ async def test_check_health(server: RemoteOpenAIServer):
 @pytest.mark.parametrize(
     "server_args",
     [
-        pytest.param(
-            ["--max-model-len", "10100"], id="default-frontend-multiprocessing"
-        ),
-        pytest.param(
-            ["--disable-frontend-multiprocessing", "--max-model-len", "10100"],
-            id="disable-frontend-multiprocessing",
-        ),
+        pytest.param(["--max-model-len", "10100"]),
     ],
     indirect=True,
 )

tests/entrypoints/instrumentator/test_metrics.py → tests/entrypoints/serve/instrumentator/test_metrics.py

@@ -50,7 +50,6 @@ def default_server_args():
     params=[
         "",
         "--enable-chunked-prefill",
-        "--disable-frontend-multiprocessing",
         f"--show-hidden-metrics-for-version={PREV_MINOR_VERSION}",
     ],
 )

tests/entrypoints/instrumentator/test_optional_middleware.py → tests/entrypoints/serve/instrumentator/test_optional_middleware.py

@@ -10,7 +10,7 @@ from http import HTTPStatus
 import pytest
 import requests
 
-from ...utils import RemoteOpenAIServer
+from tests.utils import RemoteOpenAIServer
 
 # Use a small embeddings model for faster startup and smaller memory footprint.
 # Since we are not testing any chat functionality,

tests/entrypoints/instrumentator/test_orca_metrics.py → tests/entrypoints/serve/instrumentator/test_orca_metrics.py

@@ -5,7 +5,7 @@ import openai
 import pytest
 import pytest_asyncio
 
-from ...utils import RemoteOpenAIServer
+from tests.utils import RemoteOpenAIServer
 
 # any model with a chat template should work here
 MODEL_NAME = "Qwen/Qwen3-0.6B"

tests/entrypoints/openai/test_lora_adapters.py → tests/entrypoints/serve/lora/test_lora_adapters.py

@@ -10,7 +10,7 @@ import openai  # use the official client for correctness check
 import pytest
 import pytest_asyncio
 
-from ...utils import RemoteOpenAIServer
+from tests.utils import RemoteOpenAIServer
 
 # any model with a chat template should work here
 MODEL_NAME = "Qwen/Qwen3-0.6B"

tests/entrypoints/openai/test_launch_render.py → tests/entrypoints/serve/render/test_launch_render.py

@@ -6,7 +6,7 @@ import httpx
 import pytest
 import pytest_asyncio
 
-from ...utils import RemoteLaunchRenderServer
+from tests.utils import RemoteLaunchRenderServer
 
 MODEL_NAME = "hmellor/tiny-random-LlamaForCausalLM"
 

tests/entrypoints/openai/test_tokenization.py → tests/entrypoints/serve/tokenize/test_tokenization.py

@@ -5,10 +5,9 @@ import pytest
 import pytest_asyncio
 import requests
 
+from tests.utils import RemoteOpenAIServer
 from vllm.tokenizers import get_tokenizer
 
-from ...utils import RemoteOpenAIServer
-
 # any model with a chat template should work here
 MODEL_NAME = "HuggingFaceH4/zephyr-7b-beta"
 

tests/entrypoints/openai/test_tokenization_vlm.py → tests/entrypoints/serve/tokenize/test_tokenization_vlm.py

@@ -13,7 +13,7 @@ import json
 import pytest
 import requests
 
-from ...utils import RemoteOpenAIServer
+from tests.utils import RemoteOpenAIServer
 
 MODEL_NAME = "Qwen/Qwen2.5-VL-3B-Instruct"
 

tests/evals/gpt_oss/configs/models-gfx942.txt

 # GFX942 model configurations for GPQA evaluation
 # Tests different environment variable combinations
-gpt-oss-20b-rocm-baseline.yaml
\ No newline at end of file
+gpt-oss-20b-rocm-baseline.yaml

tests/evals/gsm8k/configs/DeepSeek-R1-DP_MI325.yaml

+model_name: "deepseek-ai/DeepSeek-R1"
+accuracy_threshold: 0.95
+num_questions: 1319
+num_fewshot: 5
+startup_max_wait_seconds: 1200
+server_args: >-
+  --enforce-eager
+  --max-model-len 4096
+  --data-parallel-size 8
+  --enable-expert-parallel
+  --attention-backend=TRITON_ATTN
+  --speculative-config '{"method":"mtp","num_speculative_tokens":3}'

tests/evals/gsm8k/configs/DeepSeek-R1-TP_MI325.yaml

+model_name: "deepseek-ai/DeepSeek-R1"
+accuracy_threshold: 0.95
+num_questions: 1319
+num_fewshot: 5
+startup_max_wait_seconds: 1200
+server_args: >-
+  --enforce-eager
+  --max-model-len 4096
+  --tensor-parallel-size 8
+  --enable-expert-parallel
+  --attention-backend=TRITON_ATTN
+  --speculative-config '{"method":"mtp","num_speculative_tokens":3}'

tests/evals/gsm8k/configs/DeepSeek-V3.2-DP_MI325.yaml

+model_name: "deepseek-ai/DeepSeek-V3.2"
+accuracy_threshold: 0.95
+num_questions: 1319
+num_fewshot: 5
+startup_max_wait_seconds: 1200
+server_args: >-
+  --enforce-eager
+  --max-model-len 4096
+  --data-parallel-size 8
+  --enable-expert-parallel
+  --attention-backend=TRITON_ATTN
+  --speculative-config '{"method":"mtp","num_speculative_tokens":3}'