A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 10, 22:45 UTC
[Chore] Remove unused `noqa`s (#31263)
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
Cyrus Leung committed
Dec 24, 2025 at 21:38 UTC
aa3868ecfe65036349412ce54fa5b07cb545d836
17 files changed
+17
-26
tests/conftest.py
+1
-1
@@ -410,7 +410,7 @@ class HfRunner:
410
411
# don't put this import at the top level
412
# it will call torch.cuda.device_count()
413
- from transformers import AutoProcessor # noqa: F401
413
+ from transformers import AutoProcessor
414
415
self.processor = AutoProcessor.from_pretrained(
416
model_name,
tests/entrypoints/openai/test_async_tokenization.py
+1
-1
@@ -15,7 +15,7 @@ MODEL_NAME = "Qwen/Qwen2.5-1.5B-Instruct"
15
16
17
@pytest.fixture(scope="module")
18
-def server(): # noqa: F811
18
+def server():
19
args = [
20
# use half precision for speed and memory savings in CI environment
21
"--dtype",
tests/entrypoints/openai/test_chat.py
+1
-1
@@ -28,7 +28,7 @@ def zephyr_lora_files():
28
29
30
@pytest.fixture(scope="module")
31
-def server(zephyr_lora_files): # noqa: F811
31
+def server(zephyr_lora_files):
32
args = [
33
# use half precision for speed and memory savings in CI environment
34
"--dtype",
tests/entrypoints/openai/test_chat_with_tool_reasoning.py
+1
-1
@@ -12,7 +12,7 @@ MODEL_NAME = "Qwen/QwQ-32B"
12
13
14
@pytest.fixture(scope="module")
15
-def server(): # noqa: F811
15
+def server():
16
args = [
17
"--max-model-len",
18
"8192",
tests/entrypoints/openai/test_completion_with_function_calling.py
+2
-2
@@ -125,7 +125,7 @@ messages = [
125
126
127
@pytest.fixture(scope="module")
128
-def server(): # noqa: F811
128
+def server():
129
args = [
130
# use half precision for speed and memory savings in CI environment
131
"--dtype",
@@ -212,7 +212,7 @@ async def test_function_tool_use(
212
213
214
@pytest.fixture(scope="module")
215
-def k2_server(): # noqa: F811
215
+def k2_server():
216
args = [
217
# use half precision for speed and memory savings in CI environment
218
"--dtype",
tests/entrypoints/openai/test_default_mm_loras.py
+1
-1
@@ -23,7 +23,7 @@ ACTIVE_MM_LORA_RESPONSE = "Spoken text: The first words I spoke in the original
23
24
25
@pytest.fixture(scope="module")
26
-def multimodal_server(): # noqa: F811
26
+def multimodal_server():
27
args = [
28
# use half precision for speed and memory savings in CI environment
29
"--dtype",
tests/entrypoints/openai/test_enable_force_include_usage.py
+1
-1
@@ -8,7 +8,7 @@ from ...utils import RemoteOpenAIServer
8
9
10
@pytest.fixture(scope="module")
11
-def chat_server_with_force_include_usage(request): # noqa: F811
11
+def chat_server_with_force_include_usage(request):
12
args = [
13
# use half precision for speed and memory savings in CI environment
14
"--dtype",
tests/entrypoints/openai/test_messages.py
+1
-1
@@ -11,7 +11,7 @@ MODEL_NAME = "Qwen/Qwen3-0.6B"
11
12
13
@pytest.fixture(scope="module")
14
-def server(): # noqa: F811
14
+def server():
15
args = [
16
"--max-model-len",
17
"2048",
tests/entrypoints/openai/test_return_tokens_as_ids.py
+1
-1
@@ -37,7 +37,7 @@ def default_server_args(qwen3_lora_files):
37
38
39
@pytest.fixture(scope="module")
40
-def server_fixture(request, default_server_args): # noqa: F811
40
+def server_fixture(request, default_server_args):
41
use_server_flag = request.param
42
if use_server_flag:
43
args_with_flag = default_server_args + ["--return-tokens-as-token-ids"]
tests/models/multimodal/generation/test_qwen2_vl.py
+1
-1
@@ -267,7 +267,7 @@ def run_embedding_input_test(
267
"""Inference result should be the same between
268
original image/video input and image/video embeddings input.
269
"""
270
- from transformers import AutoProcessor # noqa: F401
270
+ from transformers import AutoProcessor
271
272
processor = AutoProcessor.from_pretrained(model)
273
tests/v1/kv_connector/unit/test_example_connector.py
+1
-1
@@ -145,7 +145,7 @@ def test_shared_storage_connector_hashes(tmp_path):
145
146
# don't put this import at the top level
147
# it will call torch.cuda.device_count()
148
- from transformers import AutoProcessor # noqa: F401
148
+ from transformers import AutoProcessor
149
150
# Create processor to handle the chat prompt
151
processor = AutoProcessor.from_pretrained(MODEL_NAME)
vllm/distributed/ec_transfer/ec_connector/example_connector.py
+1
-4
@@ -81,10 +81,7 @@ class ECExampleConnector(ECConnectorBase):
81
assert encoder_cache is not None
82
if metadata is None:
83
logger.warning(
84
- (
85
- "In connector.start_load_caches, ",
86
- "but the connector metadata is None",
87
- )
84
+ "In connector.start_load_caches, but the connector metadata is None"
85
)
86
return
87
# Load the EC for each mm data
vllm/entrypoints/serve/elastic_ep/api_router.py
+1
-1
@@ -43,7 +43,7 @@ async def scale_elastic_ep(raw_request: Request):
43
try:
44
body = await raw_request.json()
45
except json.JSONDecodeError as e:
46
- raise HTTPException(status_code=400, detail="Invalid JSON format") from e # noqa: B904
46
+ raise HTTPException(status_code=400, detail="Invalid JSON format") from e
47
48
new_data_parallel_size = body.get("new_data_parallel_size")
49
drain_timeout = body.get("drain_timeout", 120) # Default 2 minutes
vllm/model_executor/model_loader/bitsandbytes_loader.py
+1
-2
@@ -1,6 +1,5 @@
1
# SPDX-License-Identifier: Apache-2.0
2
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
3
-# ruff: noqa: SIM117
3
import fnmatch
4
import glob
5
import itertools
@@ -59,7 +58,7 @@ def is_moe_model(model: torch.nn.Module) -> bool:
58
59
60
class BitsAndBytesModelLoader(BaseModelLoader):
62
- """Model loader to load model weights with BitAndBytes quantization."""
61
+ """Model loader to load model weights with BitsAndBytes quantization."""
62
63
possible_config_file_names = ["adapter_config.json"]
64
vllm/model_executor/model_loader/runai_streamer_loader.py
-1
@@ -1,6 +1,5 @@
1
# SPDX-License-Identifier: Apache-2.0
2
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
3
-# ruff: noqa: SIM117
3
import os
4
from collections.abc import Generator
5
vllm/v1/worker/ec_connector_model_runner_mixin.py
+1
-3
@@ -6,9 +6,7 @@ Define EC connector functionality mixin for model runners.
6
7
from collections.abc import Generator
8
from contextlib import AbstractContextManager, contextmanager, nullcontext
9
-from typing import (
10
- TYPE_CHECKING, # noqa: UP035
11
-)
9
+from typing import TYPE_CHECKING
10
11
import torch
12
vllm/v1/worker/kv_connector_model_runner_mixin.py
+1
-3
@@ -7,9 +7,7 @@ Define KV connector functionality mixin for model runners.
7
import copy
8
from collections.abc import Generator
9
from contextlib import AbstractContextManager, contextmanager, nullcontext
10
-from typing import (
11
- TYPE_CHECKING, # noqa: UP035
12
-)
10
+from typing import TYPE_CHECKING
11
12
import torch
13