vllm-project/vllm

Pattern Findings

14571 matches across 21 categories. Click a row to expand file-level details.

Hyper-Verbose Identifiers8193 hits · 8109 pts

Severity	File	Line	Snippet
LOW	setup.py	52	def should_require_rust_frontend() -> bool:
LOW	setup.py	445	def fetch_metadata_for_variant(
LOW	setup.py	479	def detect_system_cuda_variant() -> str:
LOW	setup.py	533	def fetch_wheel_from_pypi_index(index_url: str, package: str = "vllm") -> str:
LOW	setup.py	688	def extract_precompiled_and_patch_package(
LOW	setup.py	799	def get_base_commit_in_main_branch() -> str:
LOW	csrc/cpu/generate_cpu_attn_dispatch.py	92	def generate_cases_for_isa_group(isa_list: list[str], include_fp8: bool = False) -> str:
LOW	csrc/quantization/machete/generate.py	334	def generate_type_option_name(kernel_types: TypeConfig):
LOW	csrc/quantization/machete/generate.py	370	def unsigned_type_with_bitwidth(num_bits):
LOW	tools/generate_versions_json.py	86	def generate_bake_native_json(args: dict[str, str]) -> dict:
LOW	tools/install_nixl_from_source_ubuntu.py	65	def install_system_dependencies():
LOW	tools/install_nixl_from_source_ubuntu.py	105	def build_and_install_prerequisites(args):
LOW	tools/pre_commit/generate_attention_backend_docs.py	1046	def _expand_flash_attn_variants(
LOW	tools/pre_commit/generate_attention_backend_docs.py	1150	def parse_cuda_priority_lists() -> dict[str, list[str]]:
LOW	tools/pre_commit/generate_attention_backend_docs.py	300	def parse_mla_prefill_registry() -> dict[str, str]:
LOW	tools/pre_commit/generate_attention_backend_docs.py	320	def parse_mla_prefill_priorities() -> dict[str, list[str]]:
LOW	tools/pre_commit/generate_attention_backend_docs.py	386	def parse_mla_prefill_backend_file(class_path: str) -> dict[str, Any] \| None:
LOW	tools/pre_commit/generate_attention_backend_docs.py	466	def parse_mla_prefill_backends() -> list[dict[str, Any]]:
LOW	tools/pre_commit/generate_attention_backend_docs.py	866	def parse_flash_attn_features() -> dict[str, dict[str, Any]]:
LOW	tools/pre_commit/generate_attention_backend_docs.py	1003	def parse_flashinfer_trtllm_features() -> dict[str, dict[str, Any]]:
LOW	tools/pre_commit/generate_attention_backend_docs.py	1098	def _expand_flashinfer_variants(
LOW	tools/pre_commit/generate_attention_backend_docs.py	1203	def _get_backends_from_return(stmts: list) -> list[str]:
LOW	tools/pre_commit/generate_attention_backend_docs.py	1490	def generate_priority_section(priorities: dict[str, list[str]]) -> str:
LOW	tools/profiler/visualize_layerwise_profile.py	75	def shorten_plot_legend_strings(legend, max_char_len: int):
LOW	tools/profiler/visualize_layerwise_profile.py	94	def attempt_to_make_names_unique(entries_and_traces):
LOW	tools/profiler/visualize_layerwise_profile.py	144	def group_trace_by_operations(trace_df: "pd.DataFrame") -> "pd.DataFrame":
LOW	tools/profiler/nsys_profile_tools/gputrc2graph.py	45	def gen_nonoverlapped_sum_from_gputrace(self, in_file, out_file):
LOW	tools/profiler/nsys_profile_tools/gputrc2graph.py	66	def sum_non_overlapping_intervals(self, df):
LOW	tools/vllm-rocm/pin_rocm_dependencies.py	20	def extract_version_from_wheel(wheel_name: str) -> str:
LOW	tools/vllm-rocm/pin_rocm_dependencies.py	40	def get_custom_wheel_versions(install_dir: str) -> dict[str, str]:
LOW	tools/vllm-rocm/pin_rocm_dependencies.py	94	def pin_dependencies_in_requirements(requirements_path: str, versions: dict[str, str]):
LOW	tests/test_sequence.py	9	def test_sequence_intermediate_tensors_equal():
LOW	tests/test_zen_cpu_platform_detection.py	35	def test_is_amd_zen_cpu_returns_false_when_cpuinfo_missing():
LOW	tests/test_version.py	36	def test_prev_minor_version_was(version_tuple, version_str, expected):
LOW	tests/test_ray_env_utils.py	36	def test_arbitrary_var_propagated(self):
LOW	tests/test_ray_env_utils.py	42	def test_worker_specific_excluded(self):
LOW	tests/test_ray_env_utils.py	50	def test_non_carry_over_blacklist(self):
LOW	tests/test_fxgraphcache_pickle_patch.py	20	def test_valueerror_converted_to_bypass(self):
LOW	tests/test_fxgraphcache_pickle_patch.py	30	def test_original_valueerror_chained(self):
LOW	tests/test_fxgraphcache_pickle_patch.py	44	def test_non_valueerror_propagates(self):
LOW	tests/test_fxgraphcache_pickle_patch.py	54	def test_normal_return_preserved(self):
LOW	tests/test_fxgraphcache_pickle_patch.py	76	def test_sentinel_attribute_set(self):
LOW	tests/test_fxgraphcache_pickle_patch.py	90	def test_patch_applied_in_current_environment():
LOW	tests/conftest.py	217	def init_test_http_connection():
LOW	tests/conftest.py	264	def should_do_global_cleanup_after_test(request) -> bool:
LOW	tests/conftest.py	738	def _hidden_states_to_seq_logprobs(
LOW	tests/conftest.py	761	def _hidden_states_to_logprobs(
LOW	tests/conftest.py	788	def generate_greedy_logprobs_limit(
LOW	tests/conftest.py	1027	def _final_steps_generate_w_logprobs(
LOW	tests/conftest.py	1137	def generate_prompt_perplexity(
LOW	tests/conftest.py	1250	def _wait_for_rocm_memory_release(self, gpu_memory_utilization: float) -> None:
LOW	tests/conftest.py	1295	def temporary_enable_log_propagate():
LOW	tests/conftest.py	1499	def pytest_collection_modifyitems(config, items):
LOW	tests/conftest.py	1516	def cli_config_file_with_model():
LOW	tests/conftest.py	1654	def clean_gpu_memory_between_tests():
LOW	tests/test_inputs.py	21	def test_preprocessor_always_mm_code_path(model_id, prompt):
LOW	tests/test_ray_env.py	54	def test_pythonhashseed_in_result(self):
LOW	tests/test_ray_env.py	118	def test_worker_specific_host_vars_are_excluded(self):
LOW	tests/test_ray_env.py	128	def test_non_carry_over_blacklist(self):
LOW	tests/test_ray_env.py	142	def test_additional_vars_passthrough(self):
8133 more matches not shown…

Decorative Section Separators1222 hits · 4246 pts

Severity	File	Line	Snippet
MEDIUM	tools/pre_commit/generate_attention_backend_docs.py	1041	# ---------------------------------------------------------------------------
MEDIUM	tools/pre_commit/generate_attention_backend_docs.py	1043	# ---------------------------------------------------------------------------
MEDIUM	tools/pre_commit/generate_attention_backend_docs.py	1145	# ---------------------------------------------------------------------------
MEDIUM	tools/pre_commit/generate_attention_backend_docs.py	1147	# ---------------------------------------------------------------------------
MEDIUM	tools/pre_commit/generate_attention_backend_docs.py	24	# ---------------------------------------------------------------------------
MEDIUM	tools/pre_commit/generate_attention_backend_docs.py	26	# ---------------------------------------------------------------------------
MEDIUM	tools/pre_commit/generate_attention_backend_docs.py	75	# ---------------------------------------------------------------------------
MEDIUM	tools/pre_commit/generate_attention_backend_docs.py	77	# ---------------------------------------------------------------------------
MEDIUM	tools/pre_commit/generate_attention_backend_docs.py	263	# ---------------------------------------------------------------------------
MEDIUM	tools/pre_commit/generate_attention_backend_docs.py	265	# ---------------------------------------------------------------------------
MEDIUM	tools/pre_commit/generate_attention_backend_docs.py	536	# ---------------------------------------------------------------------------
MEDIUM	tools/pre_commit/generate_attention_backend_docs.py	538	# ---------------------------------------------------------------------------
MEDIUM	tools/pre_commit/generate_attention_backend_docs.py	822	# ---------------------------------------------------------------------------
MEDIUM	tools/pre_commit/generate_attention_backend_docs.py	824	# ---------------------------------------------------------------------------
MEDIUM	tools/pre_commit/generate_attention_backend_docs.py	1283	# ---------------------------------------------------------------------------
MEDIUM	tools/pre_commit/generate_attention_backend_docs.py	1289	# ---------------------------------------------------------------------------
MEDIUM	tools/pre_commit/generate_attention_backend_docs.py	1382	# ---------------------------------------------------------------------------
MEDIUM	tools/pre_commit/generate_attention_backend_docs.py	1384	# ---------------------------------------------------------------------------
MEDIUM	tools/pre_commit/generate_attention_backend_docs.py	1618	# ---------------------------------------------------------------------------
MEDIUM	tools/pre_commit/generate_attention_backend_docs.py	1620	# ---------------------------------------------------------------------------
MEDIUM	docker/entrypoints/test_vllm_nonroot_entrypoint.sh	106	# -----------------------------------------------------------------------------
MEDIUM	docker/entrypoints/test_vllm_nonroot_entrypoint.sh	108	# -----------------------------------------------------------------------------
MEDIUM	docker/entrypoints/test_vllm_nonroot_entrypoint.sh	117	# -----------------------------------------------------------------------------
MEDIUM	docker/entrypoints/test_vllm_nonroot_entrypoint.sh	121	# -----------------------------------------------------------------------------
MEDIUM	docker/entrypoints/test_vllm_nonroot_entrypoint.sh	135	# -----------------------------------------------------------------------------
MEDIUM	docker/entrypoints/test_vllm_nonroot_entrypoint.sh	138	# -----------------------------------------------------------------------------
MEDIUM	docker/entrypoints/test_vllm_nonroot_entrypoint.sh	151	# -----------------------------------------------------------------------------
MEDIUM	docker/entrypoints/test_vllm_nonroot_entrypoint.sh	157	# -----------------------------------------------------------------------------
MEDIUM	docker/entrypoints/test_vllm_nonroot_entrypoint.sh	65	# -----------------------------------------------------------------------------
MEDIUM	docker/entrypoints/test_vllm_nonroot_entrypoint.sh	67	# -----------------------------------------------------------------------------
MEDIUM	docker/entrypoints/test_vllm_nonroot_entrypoint.sh	78	# -----------------------------------------------------------------------------
MEDIUM	docker/entrypoints/test_vllm_nonroot_entrypoint.sh	81	# -----------------------------------------------------------------------------
MEDIUM	docker/entrypoints/test_vllm_nonroot_entrypoint.sh	93	# -----------------------------------------------------------------------------
MEDIUM	docker/entrypoints/test_vllm_nonroot_entrypoint.sh	95	# -----------------------------------------------------------------------------
MEDIUM	docker/entrypoints/test_vllm_nonroot_entrypoint.sh	175	# -----------------------------------------------------------------------------
MEDIUM	docker/entrypoints/test_vllm_nonroot_entrypoint.sh	179	# -----------------------------------------------------------------------------
MEDIUM	docker/entrypoints/test_vllm_nonroot_entrypoint.sh	192	# -----------------------------------------------------------------------------
MEDIUM	docker/entrypoints/test_vllm_nonroot_entrypoint.sh	199	# -----------------------------------------------------------------------------
MEDIUM	docker/entrypoints/test_vllm_nonroot_entrypoint.sh	214	# -----------------------------------------------------------------------------
MEDIUM	docker/entrypoints/test_vllm_nonroot_entrypoint.sh	217	# -----------------------------------------------------------------------------
MEDIUM	docker/entrypoints/test_vllm_nonroot_entrypoint.sh	240	# -----------------------------------------------------------------------------
MEDIUM	docker/entrypoints/test_vllm_nonroot_entrypoint.sh	243	# -----------------------------------------------------------------------------
MEDIUM	tests/test_ray_env.py	11	# ---------------------------------------------------------------------------
MEDIUM	tests/test_ray_env.py	13	# ---------------------------------------------------------------------------
MEDIUM	tests/test_ray_env.py	46	# ---------------------------------------------------------------------------
MEDIUM	tests/test_ray_env.py	48	# ---------------------------------------------------------------------------
MEDIUM	tests/test_ray_env.py	61	# ---------------------------------------------------------------------------
MEDIUM	tests/test_ray_env.py	63	# ---------------------------------------------------------------------------
MEDIUM	tests/test_ray_env.py	97	# ---------------------------------------------------------------------------
MEDIUM	tests/test_ray_env.py	99	# ---------------------------------------------------------------------------
MEDIUM	tests/test_ray_env.py	133	# ---------------------------------------------------------------------------
MEDIUM	tests/test_ray_env.py	135	# ---------------------------------------------------------------------------
MEDIUM	tests/test_ray_env.py	147	# ---------------------------------------------------------------------------
MEDIUM	tests/test_ray_env.py	149	# ---------------------------------------------------------------------------
MEDIUM	tests/test_jit_monitor.py	21	# ------------------------------------------------------------------
MEDIUM	tests/test_jit_monitor.py	23	# ------------------------------------------------------------------
MEDIUM	tests/test_jit_monitor.py	39	# ------------------------------------------------------------------
MEDIUM	tests/test_jit_monitor.py	41	# ------------------------------------------------------------------
MEDIUM	tests/test_jit_monitor.py	172	# ------------------------------------------------------------------
MEDIUM	tests/test_jit_monitor.py	174	# ------------------------------------------------------------------
1162 more matches not shown…

Cross-File Repetition489 hits · 2445 pts

Severity	File	Line	Snippet
HIGH	tests/test_envs.py	0	test that callable choices raise error for invalid values.
HIGH	tests/test_envs.py	0	test that callable choices raise error for invalid values.
HIGH	tests/test_envs.py	0	test that callable choices raise error for invalid values.
HIGH	tests/test_logger.py	0	this test calls _configure_vllm_root_logger again to test custom logging config behavior, however mocks are used to ensu
HIGH	tests/test_logger.py	0	this test calls _configure_vllm_root_logger again to test custom logging config behavior, however mocks are used to ensu
HIGH	tests/test_logger.py	0	this test calls _configure_vllm_root_logger again to test custom logging config behavior, however mocks are used to ensu
HIGH	tests/test_logger.py	0	this test calls _configure_vllm_root_logger again to test custom logging config behavior, however it fails before any ch
HIGH	tests/test_logger.py	0	this test calls _configure_vllm_root_logger again to test custom logging config behavior, however it fails before any ch
HIGH	tests/test_logger.py	0	this test calls _configure_vllm_root_logger again to test custom logging config behavior, however it fails before any ch
HIGH	tests/v1/attention/test_mla_prefill_selector.py	0	clear lru cache to ensure each test case runs without caching.
HIGH	tests/kernels/attention/test_attention_selector.py	0	clear lru cache to ensure each test case runs without caching.
HIGH	tests/kernels/attention/test_mha_attn.py	0	clear lru cache to ensure each test case runs without caching.
HIGH	tests/kernels/attention/test_rocm_attention_selector.py	0	clear lru cache to ensure each test case runs without caching.
HIGH	tests/v1/logits_processors/utils.py	0	fake logit processor to support unit testing and examples
HIGH	docs/features/custom_logitsprocs.md	0	fake logit processor to support unit testing and examples
HIGH	examples/features/logits_processor/custom.py	0	fake logit processor to support unit testing and examples
HIGH	tests/v1/logits_processors/utils.py	0	the request-level logits processor masks out all logits except the token id identified by `target_token`
HIGH	docs/features/custom_logitsprocs.md	0	the request-level logits processor masks out all logits except the token id identified by `target_token`
HIGH	examples/features/logits_processor/custom_req.py	0	the request-level logits processor masks out all logits except the token id identified by `target_token`
HIGH	examples/features/logits_processor/custom_req_init.py	0	the request-level logits processor masks out all logits except the token id identified by `target_token`
HIGH	tests/v1/logits_processors/utils.py	0	example of wrapping a fake request-level logit processor to create a batch-level logits processor
HIGH	docs/features/custom_logitsprocs.md	0	example of wrapping a fake request-level logit processor to create a batch-level logits processor
HIGH	examples/features/logits_processor/custom_req.py	0	example of wrapping a fake request-level logit processor to create a batch-level logits processor
HIGH	tests/v1/logits_processors/utils.py	0	this method returns a new request-level logits processor, customized to the `target_token` value associated with a parti
HIGH	docs/features/custom_logitsprocs.md	0	this method returns a new request-level logits processor, customized to the `target_token` value associated with a parti
HIGH	examples/features/logits_processor/custom_req.py	0	this method returns a new request-level logits processor, customized to the `target_token` value associated with a parti
HIGH	…s/v1/kv_connector/nixl_integration/toy_proxy_server.py	0	lifespan context manager to handle startup and shutdown events.
HIGH	…cache/disagg_prefill_lmcache_v1/disagg_proxy_server.py	0	lifespan context manager to handle startup and shutdown events.
HIGH	…regated/mooncake_connector/mooncake_connector_proxy.py	0	lifespan context manager to handle startup and shutdown events.
HIGH	tests/tool_parsers/test_seed_oss_tool_parser.py	0	<seed:think>the user\'s current thinking budget is 512.</seed:cot_budget_reflect>\nlet me analyze the
HIGH	tests/tool_parsers/test_seed_oss_tool_parser.py	0	<seed:think>the user\'s current thinking budget is 512.</seed:cot_budget_reflect>\nlet me analyze the
HIGH	tests/tool_parsers/test_seed_oss_tool_parser.py	0	<seed:think>the user\'s current thinking budget is 512.</seed:cot_budget_reflect>\nlet me analyze the
HIGH	tests/tool_parsers/test_seed_oss_tool_parser.py	0	<seed:think>the user\'s current thinking budget is 512.</seed:cot_budget_reflect>\nlet me analyze the
HIGH	tests/tool_parsers/test_seed_oss_tool_parser.py	0	question. the user wants to know the weather in barcelona, spain. looking at the functions available,
HIGH	tests/tool_parsers/test_seed_oss_tool_parser.py	0	question. the user wants to know the weather in barcelona, spain. looking at the functions available,
HIGH	tests/tool_parsers/test_seed_oss_tool_parser.py	0	question. the user wants to know the weather in barcelona, spain. looking at the functions available,
HIGH	tests/tool_parsers/test_seed_oss_tool_parser.py	0	question. the user wants to know the weather in barcelona, spain. looking at the functions available,
HIGH	tests/tool_parsers/test_seed_oss_tool_parser.py	0	there\'s a get_weather function that can retrieve the current temperature for a given location. \n\nfirst,
HIGH	tests/tool_parsers/test_seed_oss_tool_parser.py	0	there\'s a get_weather function that can retrieve the current temperature for a given location. \n\nfirst,
HIGH	tests/tool_parsers/test_seed_oss_tool_parser.py	0	there\'s a get_weather function that can retrieve the current temperature for a given location. \n\nfirst,
HIGH	tests/tool_parsers/test_seed_oss_tool_parser.py	0	there\'s a get_weather function that can retrieve the current temperature for a given location. \n\nfirst,
HIGH	tests/tool_parsers/test_seed_oss_tool_parser.py	0	check the parameters required by get_weather: location is mandatory (needs city and country), and unit is
HIGH	tests/tool_parsers/test_seed_oss_tool_parser.py	0	check the parameters required by get_weather: location is mandatory (needs city and country), and unit is
HIGH	tests/tool_parsers/test_seed_oss_tool_parser.py	0	check the parameters required by get_weather: location is mandatory (needs city and country), and unit is
HIGH	tests/tool_parsers/test_seed_oss_tool_parser.py	0	check the parameters required by get_weather: location is mandatory (needs city and country), and unit is
HIGH	tests/tool_parsers/test_seed_oss_tool_parser.py	0	optional. the user provided "barcelona spain" as the location, which fits the required format (city,
HIGH	tests/tool_parsers/test_seed_oss_tool_parser.py	0	optional. the user provided "barcelona spain" as the location, which fits the required format (city,
HIGH	tests/tool_parsers/test_seed_oss_tool_parser.py	0	optional. the user provided "barcelona spain" as the location, which fits the required format (city,
HIGH	tests/tool_parsers/test_seed_oss_tool_parser.py	0	optional. the user provided "barcelona spain" as the location, which fits the required format (city,
HIGH	tests/tool_parsers/test_seed_oss_tool_parser.py	0	country). \n<seed:cot_budget_reflect>i have used 131 tokens, and there are 381 tokens remaining for use.
HIGH	tests/tool_parsers/test_seed_oss_tool_parser.py	0	country). \n<seed:cot_budget_reflect>i have used 131 tokens, and there are 381 tokens remaining for use.
HIGH	tests/tool_parsers/test_seed_oss_tool_parser.py	0	country). \n<seed:cot_budget_reflect>i have used 131 tokens, and there are 381 tokens remaining for use.
HIGH	tests/tool_parsers/test_seed_oss_tool_parser.py	0	country). \n<seed:cot_budget_reflect>i have used 131 tokens, and there are 381 tokens remaining for use.
HIGH	tests/tool_parsers/test_seed_oss_tool_parser.py	0	</seed:cot_budget_reflect>\n since the unit isn\'t specified, the function will default to celsius, which
HIGH	tests/tool_parsers/test_seed_oss_tool_parser.py	0	</seed:cot_budget_reflect>\n since the unit isn\'t specified, the function will default to celsius, which
HIGH	tests/tool_parsers/test_seed_oss_tool_parser.py	0	</seed:cot_budget_reflect>\n since the unit isn\'t specified, the function will default to celsius, which
HIGH	tests/tool_parsers/test_seed_oss_tool_parser.py	0	</seed:cot_budget_reflect>\n since the unit isn\'t specified, the function will default to celsius, which
HIGH	tests/tool_parsers/test_seed_oss_tool_parser.py	0	is fine. \n\nthere\'s no need to ask for more information because the location is clear. so i should call
HIGH	tests/tool_parsers/test_seed_oss_tool_parser.py	0	is fine. \n\nthere\'s no need to ask for more information because the location is clear. so i should call
HIGH	tests/tool_parsers/test_seed_oss_tool_parser.py	0	is fine. \n\nthere\'s no need to ask for more information because the location is clear. so i should call
429 more matches not shown…

Unused Imports1159 hits · 1088 pts

Severity	File	Line
LOW	…c/cutlass_extensions/vllm_cutlass_library_extension.py	6
LOW	tests/conftest.py	75
LOW	tests/conftest.py	75
LOW	tests/conftest.py	76
LOW	tests/utils.py	63
LOW	tests/utils.py	1951
LOW	tests/v1/attention/test_mla_backends.py	69
LOW	tests/v1/attention/test_attention_backends.py	46
LOW	tests/v1/logits_processors/utils.py	13
LOW	tests/v1/cudagraph/test_cudagraph_mode.py	38
LOW	tests/v1/cudagraph/test_breakable_cudagraph.py	7
LOW	tests/v1/kv_connector/unit/test_tp_mapping.py	10
LOW	tests/v1/kv_connector/unit/test_hf3fs_client.py	17
LOW	tests/v1/kv_connector/unit/test_hf3fs_client.py	17
LOW	tests/v1/kv_connector/unit/test_hf3fs_client.py	17
LOW	tests/v1/kv_connector/unit/test_hf3fs_client.py	17
LOW	tests/v1/kv_connector/unit/test_hf3fs_client.py	17
LOW	…/v1/kv_connector/unit/offloading_connector/conftest.py	3
LOW	tests/v1/spec_decode/test_backup_token_async_spec.py	9
LOW	tests/v1/sample/test_topk_topp_sampler.py	27
LOW	tests/v1/engine/conftest.py	20
LOW	tests/v1/engine/conftest.py	20
LOW	tests/v1/simple_kv_offload/test_scheduler.py	5
LOW	tests/tool_use/test_gemma4_responses_adjust_request.py	25
LOW	tests/renderers/test_chat_utils_prompt_embeds.py	6
LOW	tests/kernels/core/test_vit_fp8_attn.py	20
LOW	tests/kernels/core/test_fused_q_kv_rmsnorm.py	11
LOW	tests/kernels/ir/test_ir_ops.py	11
LOW	tests/kernels/ir/test_layernorm.py	7
LOW	tests/kernels/mamba/test_ssu_dispatch.py	24
LOW	tests/kernels/moe/test_moe.py	18
LOW	tests/distributed/test_eplb_spec_decode.py	3
LOW	…add_dummy_platform/vllm_add_dummy_platform/__init__.py	10
LOW	tests/cuda/scripts/check_device_count_respects_env.py	14
LOW	tests/model_executor/test_oink_integration.py	37
LOW	tests/models/language/pooling/embed_utils.py	8
LOW	tests/models/language/pooling/test_reward.py	18
LOW	tests/models/multimodal/generation/test_pixtral.py	24
LOW	tests/vllm_test_utils/vllm_test_utils/__init__.py	8
LOW	tests/vllm_test_utils/vllm_test_utils/__init__.py	8
LOW	tests/vllm_test_utils/vllm_test_utils/__init__.py	9
LOW	tests/vllm_test_utils/vllm_test_utils/__init__.py	9
LOW	tests/compile/test_compile_ranges.py	10
LOW	tests/compile/test_config.py	32
LOW	tests/compile/test_structured_logging.py	10
LOW	tests/compile/test_graph_partition.py	21
LOW	tests/compile/test_decorator.py	21
LOW	tests/compile/fullgraph/test_toy_llama.py	34
LOW	tests/compile/fullgraph/test_multiple_graphs.py	28
LOW	tests/compile/passes/ir/test_lowering.py	7
LOW	…ranscription/test_transcription_inter_chunk_spacing.py	11
LOW	tests/entrypoints/openai/test_dp_supervisor.py	17
LOW	tests/entrypoints/openai/responses/conftest.py	3
LOW	tests/entrypoints/openai/responses/test_mcp_tools.py	5
LOW	tests/entrypoints/openai/responses/test_harmony.py	5
LOW	tests/standalone_tests/lazy_imports.py	22
LOW	vllm/__init__.py	7
LOW	vllm/__init__.py	7
LOW	vllm/__init__.py	14
LOW	vllm/__init__.py	42
1099 more matches not shown…

Self-Referential Comments322 hits · 1028 pts

Severity	File	Line	Snippet
MEDIUM	tools/install_deepgemm.sh	83	# Create a temporary directory for the build
MEDIUM	tools/report_build_time_ninja.py	201	# Create a list that is in order by time stamp and has entries for the
MEDIUM	tools/pre_commit/update-dockerfile-graph.sh	26	# Define the target file path
MEDIUM	tests/conftest.py	551	# Create a copy to avoid modifying the original dict
MEDIUM	tests/utils.py	727	# Create a dedicated process group so we can kill
MEDIUM	tests/utils.py	1493	# Create a unique temporary file to store exception info from child
MEDIUM	tests/test_access_log_filter.py	259	# Create a logger with our filter (simulating uvicorn.access)
MEDIUM	tests/test_access_log_filter.py	266	# Create a custom handler that tracks messages
MEDIUM	tests/test_config.py	694	# Create a new mock and run the method with the same S3 URL
MEDIUM	tests/test_logger.py	274	# Create a mock logger to capture log calls
MEDIUM	tests/v1/test_tensor_ipc_queue.py	193	# Create a CPU tensor
MEDIUM	tests/v1/test_tensor_ipc_queue.py	511	# Create a CPU tensor
MEDIUM	tests/v1/test_tensor_ipc_queue.py	642	# Create a CPU tensor
MEDIUM	tests/v1/test_tensor_ipc_queue.py	905	# Create a tensor queue
MEDIUM	tests/v1/test_serial_utils.py	189	# Create a sample Python object
MEDIUM	tests/v1/test_serial_utils.py	207	# Create a sample tensor
MEDIUM	tests/v1/test_serial_utils.py	227	# Create a sample numpy array
MEDIUM	tests/v1/test_serial_utils.py	313	# Create a request with a non-multimodal tensor
MEDIUM	tests/v1/test_serial_utils.py	354	# Create a request with None for the tensor field
MEDIUM	tests/v1/kv_offload/test_file_mapper.py	43	# Create a copy of the mock config to avoid modifying the global one
MEDIUM	tests/v1/metrics/test_ray_metrics.py	58	# Create the actor and call the async method
MEDIUM	tests/v1/attention/test_mla_backends.py	279	# Create a realistic slot mapping that corresponds to the block table
MEDIUM	tests/v1/attention/test_mla_backends.py	1198	# Create a summary for the single-line failure message
MEDIUM	tests/v1/attention/test_attention_backends.py	187	# Create a realistic slot mapping that corresponds to the block table
MEDIUM	tests/v1/logits_processors/test_correctness.py	807	# Define a shuffled batch of requests which individually use a different
MEDIUM	tests/v1/logits_processors/test_custom_offline.py	28	# Create a mixture of requests which do and don't utilize the dummy logitproc
MEDIUM	tests/v1/logits_processors/test_custom_offline.py	63	# Create a vLLM instance and load custom logitproc
MEDIUM	tests/v1/logits_processors/test_custom_offline.py	70	# Create a reference vLLM instance without custom logitproc
MEDIUM	tests/v1/core/test_kv_cache_utils.py	252	# Create a list of KVCacheBlock objects
MEDIUM	tests/v1/core/test_kv_cache_utils.py	255	# Create a FreeKVCacheBlockQueue with these blocks
MEDIUM	tests/v1/core/test_kv_cache_utils.py	420	# Create a list of KVCacheBlock objects
MEDIUM	tests/v1/core/test_kv_cache_utils.py	423	# Create a FreeKVCacheBlockQueue with these blocks
MEDIUM	tests/v1/core/test_kv_cache_utils.py	298	# Create an empty FreeKVCacheBlockQueue with these blocks
MEDIUM	tests/v1/core/test_kv_cache_utils.py	346	# Create an empty FreeKVCacheBlockQueue
MEDIUM	tests/v1/core/test_kv_cache_utils.py	363	# Create an empty FreeKVCacheBlockQueue with these blocks
MEDIUM	tests/v1/core/test_kv_cache_utils.py	1315	# Create a VllmConfig
MEDIUM	tests/v1/core/test_kv_cache_utils.py	1351	# Create a VllmConfig
MEDIUM	tests/v1/core/test_scheduler.py	2638	# Create a request and schedule it
MEDIUM	tests/v1/core/test_scheduler.py	2665	# Create a high priority request and schedule it
MEDIUM	tests/v1/core/test_scheduler.py	3424	# Create a request and schedule it (and to be preempted)
MEDIUM	tests/v1/core/test_scheduler.py	3474	# Create a high priority request and schedule it
MEDIUM	tests/v1/core/test_scheduler.py	4263	# Create a text-only request (no mm_features).
MEDIUM	tests/v1/cudagraph/test_cudagraph_dispatch.py	55	# Create a real LoRAConfig with specialize_active_lora enabled
MEDIUM	tests/v1/kv_connector/unit/test_nixl_connector.py	1912	# Create a request that triggers do_remote_decode so that
MEDIUM	tests/v1/kv_connector/unit/test_lmcache_connector.py	216	# Create a mock object that is not LMCacheKVEvents
MEDIUM	tests/v1/kv_connector/unit/test_moriio_connector.py	173	# Define a fake remote engine id for testing
MEDIUM	…ts/v1/kv_connector/unit/test_decode_bench_connector.py	145	# Create a request with multiple blocks worth of tokens
MEDIUM	…ts/v1/kv_connector/unit/test_decode_bench_connector.py	189	# Create a request
MEDIUM	…ts/v1/kv_connector/unit/test_decode_bench_connector.py	211	# Create a request with just 1 token
MEDIUM	…ts/v1/kv_connector/unit/test_decode_bench_connector.py	229	# Create a request with 2 tokens
MEDIUM	…ts/v1/kv_connector/unit/test_decode_bench_connector.py	255	# Create a request with many blocks
MEDIUM	…ts/v1/kv_connector/unit/test_decode_bench_connector.py	338	# Create a request that doesn't align to block boundaries
MEDIUM	tests/v1/kv_connector/unit/test_nixl_connector_hma.py	75	# Create a mock worker with just the required attributes
MEDIUM	tests/v1/kv_connector/unit/test_example_connector.py	146	# Create the LLM instance
MEDIUM	…r/extract_hidden_states_integration/test_extraction.py	48	# Create a minimal Llama config with small dimensions
MEDIUM	…r/extract_hidden_states_integration/test_extraction.py	63	# Create a simple tokenizer
MEDIUM	tests/v1/determinism/test_batch_invariance.py	102	# Create a batch of size `max_batch_size` and insert the needle at
MEDIUM	tests/v1/distributed/test_external_lb_dp.py	154	# Create a client for each server
MEDIUM	tests/v1/distributed/test_hybrid_lb_dp.py	182	# Create a client for each node (each node has its own API endpoint)
MEDIUM	tests/v1/streaming_input/test_async_llm_streaming.py	20	# Create a minimal mock without initializing the full engine
262 more matches not shown…

Deep Nesting1055 hits · 966 pts

Severity	File	Line
LOW	use_existing_torch.py	21
LOW	setup.py	943
LOW	setup.py	987
LOW	setup.py	173
LOW	setup.py	589
LOW	setup.py	688
LOW	csrc/cpu/generate_cpu_attn_dispatch.py	92
LOW	csrc/quantization/marlin/generate_kernels.py	173
LOW	csrc/moe/marlin_moe_wna16/generate_kernels.py	173
LOW	tools/report_build_time_ninja.py	151
LOW	tools/pre_commit/generate_attention_backend_docs.py	115
LOW	tools/pre_commit/generate_attention_backend_docs.py	140
LOW	tools/pre_commit/generate_attention_backend_docs.py	161
LOW	tools/pre_commit/generate_attention_backend_docs.py	206
LOW	tools/pre_commit/generate_attention_backend_docs.py	320
LOW	tools/pre_commit/generate_attention_backend_docs.py	386
LOW	tools/pre_commit/generate_attention_backend_docs.py	550
LOW	tools/pre_commit/generate_attention_backend_docs.py	594
LOW	tools/pre_commit/generate_attention_backend_docs.py	866
LOW	tools/pre_commit/generate_attention_backend_docs.py	1150
LOW	tools/pre_commit/generate_attention_backend_docs.py	1203
LOW	tools/pre_commit/validate_config.py	73
LOW	tools/pre_commit/check_boolean_context_manager.py	21
LOW	tools/pre_commit/check_spdx_header.py	27
LOW	tools/pre_commit/check_spdx_header.py	65
LOW	tools/pre_commit/check_spdx_header.py	109
LOW	tools/vllm-rocm/pin_rocm_dependencies.py	40
LOW	tools/vllm-rocm/pin_rocm_dependencies.py	94
LOW	tests/conftest.py	525
LOW	tests/conftest.py	941
LOW	tests/utils.py	1150
LOW	tests/utils.py	1484
LOW	tests/utils.py	1610
LOW	tests/utils.py	535
LOW	tests/utils.py	1490
LOW	tests/utils.py	1633
LOW	tests/v1/utils.py	12
LOW	tests/v1/kv_offload/cpu/test_gpu_worker.py	223
LOW	tests/v1/attention/test_mla_backends.py	721
LOW	tests/v1/attention/test_sparse_mla_backends.py	531
LOW	tests/v1/logits_processors/test_correctness.py	304
LOW	tests/v1/logits_processors/test_correctness.py	456
LOW	tests/v1/core/test_scheduler.py	1923
LOW	tests/v1/core/utils.py	176
LOW	tests/v1/kv_connector/unit/test_nixl_connector.py	2561
LOW	tests/v1/kv_connector/unit/utils.py	317
LOW	tests/v1/kv_connector/unit/test_mooncake_connector.py	315
LOW	tests/v1/kv_connector/unit/test_mooncake_connector.py	704
LOW	tests/v1/determinism/test_batch_invariance.py	27
LOW	tests/v1/determinism/test_batch_invariance.py	154
LOW	tests/v1/determinism/test_batch_invariance.py	645
LOW	tests/v1/determinism/test_nvfp4_batch_invariant.py	45
LOW	tests/v1/spec_decode/test_eagle.py	48
LOW	tests/v1/spec_decode/test_acceptance_length.py	178
LOW	tests/v1/spec_decode/test_acceptance_length.py	228
LOW	tests/v1/sample/test_logprobs.py	129
LOW	tests/v1/sample/test_logprobs.py	494
LOW	tests/v1/sample/test_logprobs.py	866
LOW	tests/v1/sample/test_logprobs.py	910
LOW	tests/v1/sample/test_logprobs.py	956
995 more matches not shown…

Over-Commented Block785 hits · 716 pts

Severity	File	Line	Snippet
LOW	CMakeLists.txt	1	cmake_minimum_required(VERSION 3.26)
LOW	CMakeLists.txt	41	set(PYTHON_SUPPORTED_VERSIONS "3.10" "3.11" "3.12" "3.13" "3.14")
LOW	CMakeLists.txt	1081
LOW	csrc/torch_utils.h	1	#pragma once
LOW	csrc/torch_bindings.cpp	1	// Provides torch::Tensor for ops.h (previously included transitively via
LOW	csrc/torch_bindings.cpp	61	"Tensor q_in, Tensor kv, Tensor! k_cache, "
LOW	csrc/cumem_allocator_compat.h	101	} // extern "C"
LOW	csrc/cub_helpers.h	1	#pragma once
LOW	csrc/launch_bounds_utils.h	1	#pragma once
LOW	csrc/cumem_allocator.cpp	1	// A CUDAPluggableAllocator based on cumem* APIs.
LOW	csrc/cuda_compat.h	41	#define VLLM_LDG(arg) __ldg(arg)
LOW	csrc/cuda_compat.h	61	#endif
LOW	csrc/cuda_utils.h	1	#pragma once
LOW	csrc/spinloop.cpp	1	#include <Python.h>
LOW	csrc/attention/attention_dtypes.h	1	#pragma once
LOW	csrc/core/scalar_type.hpp	1	#pragma once
LOW	csrc/core/registration.h	1	#pragma once
LOW	csrc/cpu/cpu_attn_neon.hpp	1	#ifndef CPU_ATTN_NEON_HPP
LOW	csrc/cpu/utils.cpp	1	#ifndef VLLM_NUMA_DISABLED
LOW	csrc/cpu/cpu_fused_moe.cpp	1	#include "cpu/cpu_types.hpp"
LOW	csrc/cpu/cpu_types.hpp	1	#ifndef CPU_TYPES_HPP
LOW	csrc/cpu/cpu_types.hpp	21	#include "cpu_types_scalar.hpp"
LOW	csrc/cpu/cpu_types_riscv.hpp	1	#ifndef CPU_TYPES_RISCV_HPP
LOW	csrc/cpu/cpu_attn_impl.hpp	901	static constexpr int64_t head_dim = attention_impl_t::HeadDim;
LOW	csrc/cpu/cpu_attn_impl.hpp	921	// BlockSizeAlignment
LOW	csrc/cpu/cpu_attn_impl.hpp	1741	blocksize_alignment);
LOW	csrc/cpu/cpu_attn_impl.hpp	1761	float* curr_partial_q_buffer =
LOW	csrc/cpu/cpu_attn_vxe.hpp	381	} // namespace cpu_attention
LOW	csrc/cpu/generate_cpu_attn_dispatch.py	141	#ifdef CPU_CAPABILITY_AMXBF16
LOW	csrc/cpu/cpu_arch_macros.h	61	#endif
LOW	csrc/cpu/cpu_arch_macros.h	161	#include <riscv_vector.h>
LOW	csrc/cpu/cpu_attn_rvv.hpp	1	// SPDX-License-Identifier: Apache-2.0
LOW	csrc/cpu/cpu_attn_fp8.hpp	1	// SPDX-License-Identifier: Apache-2.0
LOW	csrc/cpu/cpu_types_vxe.hpp	1
LOW	csrc/cpu/cpu_types_vxe.hpp	21
LOW	csrc/cpu/cpu_types_x86.hpp	1
LOW	csrc/cpu/shm.cpp	1	#include "cpu/cpu_types.hpp"
LOW	csrc/cpu/cpu_types_riscv_impl.hpp	1	#ifndef CPU_TYPES_RISCV_IMPL_HPP
LOW	csrc/cpu/cpu_types_riscv_impl.hpp	921	#define CPU_KERNEL_GUARD_IN(NAME)
LOW	csrc/cpu/cpu_types_riscv_defs.hpp	1	#ifndef CPU_TYPES_RISCV_DEFS_HPP
LOW	csrc/cpu/cpu_types_riscv_defs.hpp	21	#define LMUL_256 m1
LOW	csrc/cpu/cpu_types_arm.hpp	21	#define VLLM_DISPATCH_CASE_FLOATING_TYPES(...) \
LOW	csrc/cpu/cpu_attn_vsx.hpp	1	// SPDX-License-Identifier: Apache-2.0
LOW	csrc/cpu/sgl-kernels/gemm.cpp	81	constexpr int BLOCK_N = block_size_n();
LOW	csrc/cpu/sgl-kernels/fla.cpp	1	// Adapted from
LOW	csrc/cpu/sgl-kernels/gemm.h	1	// Adapted from
LOW	csrc/cpu/sgl-kernels/vec.h	401	__m512i vec_zero = _mm512_setzero_epi32();
LOW	csrc/cpu/sgl-kernels/moe.cpp	1	// Adapted from
LOW	csrc/cpu/sgl-kernels/moe.cpp	21	// allocates 2 intermediate_caches instead of 3
LOW	csrc/cpu/sgl-kernels/moe.cpp	1261	// unlike triton kernel, we fuse silu with gemm1 so only need 2 intermediate_caches:
LOW	csrc/cpu/sgl-kernels/moe_int4.cpp	61	// num_threads * BLOCK_M * K +
LOW	csrc/cpu/sgl-kernels/common.h	1	// Adapted from
LOW	csrc/cpu/sgl-kernels/common.h	201	}
LOW	csrc/cpu/sgl-kernels/common.h	281	return std::max(1, (actual_nth >> 1) * 2);
LOW	csrc/cpu/sgl-kernels/conv.cpp	141	Unroll<ROWS * COLS>{}(loadb);
LOW	csrc/cpu/sgl-kernels/conv.cpp	641	seqlen,
LOW	csrc/libtorch_stable/torch_utils.h	1	#pragma once
LOW	csrc/libtorch_stable/torch_bindings.cpp	1	#include "ops.h"
LOW	…e/attention/mla/cutlass_sm100_mla/device/sm100_mla.hpp	41
LOW	csrc/libtorch_stable/mamba/selective_scan.h	1	/******************************************************************************
725 more matches not shown…

Excessive Try-Catch Wrapping624 hits · 677 pts

Severity	File	Line	Snippet
MEDIUM	setup.py	110	def find_tcmalloc() -> Path \| None:
MEDIUM	setup.py	799	def get_base_commit_in_main_branch() -> str:
LOW	setup.py	118	except Exception:
LOW	setup.py	207	except Exception as e:
LOW	setup.py	497	except Exception:
LOW	setup.py	508	except Exception:
LOW	setup.py	635	except Exception as e:
LOW	setup.py	853	except Exception as err:
LOW	setup.py	924	except Exception:
MEDIUM	tools/generate_cmake_presets.py	168	print(f"Error writing file: {e}")
LOW	tools/install_nixl_from_source_ubuntu.py	30	except Exception:
LOW	tools/pre_commit/generate_attention_backend_docs.py	311	except Exception:
LOW	tools/pre_commit/generate_attention_backend_docs.py	332	except Exception:
LOW	tools/pre_commit/generate_attention_backend_docs.py	401	except Exception:
LOW	tools/pre_commit/generate_attention_backend_docs.py	743	except Exception:
LOW	tools/pre_commit/generate_attention_backend_docs.py	757	except Exception as e:
LOW	tools/pre_commit/generate_attention_backend_docs.py	840	except Exception:
LOW	tools/pre_commit/generate_attention_backend_docs.py	877	except Exception:
LOW	tools/pre_commit/generate_attention_backend_docs.py	1014	except Exception:
LOW	tools/pre_commit/generate_attention_backend_docs.py	1171	except Exception:
LOW	tools/profiler/nsys_profile_tools/gputrc2graph.py	239	except Exception:
LOW	tools/vllm-rocm/pin_rocm_dependencies.py	79	except Exception as e:
LOW	tests/conftest.py	1281	except Exception:
LOW	tests/conftest.py	1543	except Exception as e:
MEDIUM	tests/utils.py	117	def _nvml():
LOW	tests/utils.py	81	except Exception as e:
LOW	tests/utils.py	274	except Exception:
LOW	tests/utils.py	555	except Exception as e:
LOW	tests/utils.py	641	except Exception:
LOW	tests/utils.py	1522	except Exception as e:
LOW	tests/utils.py	1545	except Exception:
LOW	tests/utils.py	1739	except Exception as e:
LOW	tests/utils.py	1954	except Exception:
LOW	tests/v1/test_tensor_ipc_queue.py	83	except Exception as e:
LOW	tests/v1/test_tensor_ipc_queue.py	120	except Exception as e:
LOW	tests/v1/test_tensor_ipc_queue.py	308	except Exception as e:
LOW	tests/v1/test_tensor_ipc_queue.py	409	except Exception as e:
LOW	tests/v1/test_tensor_ipc_queue.py	443	except Exception as e:
LOW	tests/v1/test_tensor_ipc_queue.py	532	except Exception as e:
LOW	tests/v1/test_tensor_ipc_queue.py	570	except Exception as e:
LOW	tests/v1/test_tensor_ipc_queue.py	721	except Exception as e:
LOW	tests/v1/utils.py	64	except Exception as e:
LOW	tests/v1/kv_offload/cpu/test_shared_offload_region.py	164	except Exception as e:
LOW	tests/v1/kv_offload/cpu/test_shared_offload_region.py	123	except Exception as e:
LOW	tests/v1/shutdown/test_forward_error.py	81	except Exception as e:
LOW	tests/v1/shutdown/test_processor_error.py	39	except Exception as e:
LOW	tests/v1/kv_connector/unit/test_hf3fs_client.py	28	except Exception:
LOW	tests/v1/kv_connector/unit/test_multi_connector.py	383	except Exception as e:
LOW	tests/v1/kv_connector/unit/utils.py	352	except Exception as e:
LOW	…/v1/kv_connector/unit/test_mooncake_store_connector.py	575	except Exception:
LOW	…/v1/kv_connector/unit/test_mooncake_store_connector.py	612	except Exception:
LOW	tests/v1/kv_connector/unit/test_rixl_gpu_mem_diag.py	37	except Exception:
LOW	…s/v1/kv_connector/nixl_integration/toy_proxy_server.py	253	except Exception as e:
MEDIUM	…s/v1/kv_connector/nixl_integration/toy_proxy_server.py	258	print(f"Error occurred in disagg prefill proxy server - {api} endpoint")
MEDIUM	…/kv_connector/nixl_integration/test_disagg_accuracy.py	159	print(f"Error writing to file: {e}")
MEDIUM	…/kv_connector/nixl_integration/test_disagg_accuracy.py	168	print(f"Error writing to file: {e}")
LOW	tests/v1/determinism/test_online_batch_invariance.py	42	except Exception as e: # pragma: no cover
LOW	tests/v1/distributed/test_external_lb_dp.py	91	except Exception as e:
LOW	tests/v1/distributed/test_external_lb_dp.py	118	except Exception as e:
MEDIUM	tests/v1/distributed/test_external_lb_dp.py	119	print(f"Error stopping servers: {e}")
564 more matches not shown…

Redundant / Tautological Comments312 hits · 482 pts

Severity	File	Line	Snippet
LOW	setup.py	821	# Check if the upstream_main_commit exists in the local repo
LOW	tools/install_torchcodec_rocm.sh	15	# Check if torchcodec is already installed and working
LOW	tools/pre_commit/generate_attention_backend_docs.py	348	# Check if it's a capability.major == 10 check (Blackwell)
LOW	tools/pre_commit/generate_attention_backend_docs.py	766	# Check if this is an MLA backend by parent class or naming
LOW	tools/pre_commit/generate_attention_backend_docs.py	1188	# Check if this is the "if use_mla:" branch
LOW	tools/pre_commit/check_forbidden_imports.py	99	# Check if it's allowed
LOW	tools/pre_commit/update-dockerfile-graph.sh	10	# Check if docker/Dockerfile is among the provided files
LOW	tools/pre_commit/update-dockerfile-graph.sh	14	# Check if Docker is installed and running
LOW	tools/pre_commit/update-dockerfile-graph.sh	71	# Check if the graph has changed
LOW	tools/vllm-rocm/pin_rocm_dependencies.py	148	# Check if this line is for one of our custom packages
LOW	tests/conftest.py	385	# Set this to avoid hanging issue
LOW	tests/conftest.py	897	# Set this to avoid hanging issue
LOW	tests/utils.py	503	os.kill(spid, 0) # Check if still alive
LOW	tests/test_config.py	478	# Check if LONGCHAT_ROPE_PARAMETERS entries are in longchat_model_config
LOW	tests/test_logger.py	358	# Set max_log_len to 10
LOW	tests/v1/attention/test_mla_backends.py	830	# Set num_speculative_tokens to query_len - 1
LOW	tests/v1/attention/test_sparse_mla_backends.py	611	# Set some to -1 to test masking
LOW	tests/v1/attention/test_sparse_mla_backends.py	615	# Set some to out of bounds
LOW	tests/v1/attention/test_sparse_mla_backends.py	671	# Set some to -1 to test masking
LOW	tests/v1/attention/test_sparse_mla_backends.py	675	# Set some to out of bounds
LOW	tests/v1/core/test_scheduler.py	1987	# Verify if position length is identical
LOW	tests/v1/core/test_scheduler.py	2883	# Check if scheduled_encoder_inputs is empty as expected
LOW	tests/v1/core/test_scheduler.py	3235	# Set up to test different encoder cache existence scenario after preemption
LOW	tests/v1/core/test_scheduler.py	3565	# Set up to test different encoder cache existence scenario after preemption
LOW	tests/v1/core/utils.py	231	# Verify if position length is identical
LOW	…extract_hidden_states_integration/predictable_llama.py	81	# Check if we need auxiliary hidden states
LOW	…nnector/nixl_integration/config_sweep_accuracy_test.sh	97	# Check if cross-layers is enabled (non-empty)
LOW	tests/v1/determinism/test_batch_invariance.py	279	# Check if tokens match first
LOW	tests/v1/determinism/test_batch_invariance.py	558	# Check if tokens match first
LOW	tests/v1/determinism/test_batch_invariance.py	787	# Check if tokens match
LOW	tests/v1/determinism/test_batch_invariance.py	805	# Check if logprobs match bitwise
LOW	tests/v1/spec_decode/test_acceptance_length.py	102	# Check if get_valid_backends is actually defined in the platform class
LOW	…1/ec_connector/integration/run_epd_correctness_test.sh	25	# Set 1 to use multimodal prompts; else to use text-only
LOW	…ts/v1/ec_connector/integration/test_epd_correctness.py	218	# Check if server is ready
LOW	tests/v1/e2e/spec_decode/test_async_spec_decode.py	34	# Increment counter
LOW	tests/v1/engine/test_engine_core_client.py	284	# Check if all request IDs in outputs have finished
LOW	tests/v1/engine/utils.py	131	# Check if the sampled_token_id occurs in choice_tensor[1:]
LOW	tests/utils_/test_network_utils.py	79	# Check if IPv6 is supported by trying to create an IPv6 socket
LOW	tests/tool_parsers/test_minimax_tool_parser.py	494	# Check if function name is sent (should happen only once)
LOW	tests/tool_parsers/test_minimax_tool_parser.py	500	# Check if arguments are sent incrementally
LOW	tests/tool_parsers/test_mistral_tool_parser.py	135	# Check if the slice from the current position matches the target sequence
LOW	tests/kernels/moe/test_cpu_int4_moe.py	17	# Check if the dynamic_4bit_int_moe op is available
LOW	tests/kernels/moe/test_cpu_int4_moe.py	21	# Check if KleidiAI ops are available
LOW	tests/kernels/moe/test_moe_layer.py	1791	# Check if enough GPUs available
LOW	tests/kernels/moe/test_rocm_aiter_topk.py	26	# Check if aiter package is installed
LOW	tests/kernels/moe/test_rocm_aiter_topk.py	35	# Check if the op exists in torch.ops.vllm
LOW	tests/kernels/moe/test_rocm_aiter_topk.py	38	# Check if the op is callable
LOW	tests/kernels/moe/test_rocm_aiter_topk.py	44	# Check if the op exists in torch.ops.vllm
LOW	tests/kernels/moe/test_rocm_aiter_topk.py	47	# Check if the op is callable
LOW	tests/evals/gsm8k/gsm8k_eval.py	297	# Print results to terminal
LOW	tests/distributed/test_eplb_execute.py	150	# Check if the weights are correct
LOW	tests/model_executor/test_eagle_quantization.py	112	# Check if get_cache_scale is called and returns expected value
LOW	tests/model_executor/test_qwen3_omni.py	29	# Check if it's a special token that should be compressed
LOW	…model_loader/runai_streamer_loader/test_runai_utils.py	56	# Read the file in chunks to handle large files efficiently
LOW	tests/models/multimodal/generation/test_maverick.py	59	# Print the outputs
LOW	tests/quantization/test_gptq_v2.py	43	# Check if gptq_v2 format is correctly loaded
LOW	tests/quantization/test_gptq_v2.py	105	# Print the output sequences if failed
LOW	tests/compile/test_config.py	50	# Check if get_raw_stream exists in builtins
LOW	tests/compile/fusions_e2e/conftest.py	44	# Print the outputs.
LOW	tests/compile/fullgraph/test_full_graph.py	251	# Print the outputs.
252 more matches not shown…

AI Slop Vocabulary113 hits · 283 pts

Severity	File	Line	Snippet
LOW	csrc/quantization/machete/generate.py	435	# For now, we can just use the first accumulator type seen since
MEDIUM	docker/entrypoints/test_vllm_nonroot_entrypoint.sh	145	# More robust: count lines matching our UID.
MEDIUM	tests/v1/logits_processors/test_custom_offline.py	28	# Create a mixture of requests which do and don't utilize the dummy logitproc
LOW	tests/v1/cudagraph/test_breakable_cudagraph.py	235	# Outside capture: decorator should just call through.
LOW	tests/tool_parsers/test_mistral_tool_parser.py	141	# Otherwise, just add the current token and move to the next one
MEDIUM	tests/tool_parsers/test_glm4_moe_tool_parser.py	295	# depending on how robust we want the parsing to be
MEDIUM	tests/renderers/test_sparse_tensor_validation.py	58	# explicitly so this fixture is robust to process-wide invariant-check state
MEDIUM	tests/kernels/mamba/test_mamba_mixer2.py	106	# - utilize mock patching to disable TP when
MEDIUM	tests/distributed/test_pynccl.py	382	# Essentially this is an all-gather operation.
MEDIUM	tests/distributed/test_context_parallel.py	42	# .buildkite/lm-eval-harness/configs/DeepSeek-V2-Lite-Chat.yaml
MEDIUM	tests/distributed/test_context_parallel.py	44	# .buildkite/lm-eval-harness/configs/Qwen2.5-1.5B-Instruct.yaml
LOW	tests/model_executor/test_qwen3_omni.py	41	# Regular token, just add it
LOW	vllm/env_override.py	356	# functions just return True.
LOW	vllm/v1/attention/backends/flash_attn.py	291	# but for now just set it to `UNIFORM_BATCH` to get use to drop down
LOW	vllm/v1/attention/backends/utils.py	263	# then we can simply use a cdiv for the rest.
MEDIUM	vllm/v1/attention/ops/triton_decode_attention.py	362	# explicitly facilitate overlapping load/compute
MEDIUM	vllm/v1/core/encoder_cache_manager.py	321	# utilize the cache and this class will fold into EncoderCacheManager, as
LOW	vllm/v1/spec_decode/llm_base_proposer.py	1661	# Therefore, we can just return the logits.
MEDIUM	vllm/v1/worker/gpu_worker.py	638	# CUDAGraph memory size and may not utilize all gpu memory.
MEDIUM	vllm/v1/worker/gpu_worker.py	888	# Generate the trace name by combining prefix with comprehensive rank suffix
LOW	vllm/v1/worker/gpu/model_runner.py	1175	# For piecewise and eager mode, just call model().
LOW	vllm/v1/engine/parallel_sampling.py	116	# If streaming, just return the current output
MEDIUM	vllm/tool_parsers/llama_tool_parser.py	262	# re-set stuff pertaining to progress in the current tool
MEDIUM	vllm/tool_parsers/jamba_tool_parser.py	218	# re-set stuff pertaining to progress in the current tool
MEDIUM	vllm/tool_parsers/granite_tool_parser.py	186	# re-set stuff pertaining to progress in the current tool
LOW	vllm/tool_parsers/xlam_tool_parser.py	558	# If we encounter an error, just return the delta text as regular content
MEDIUM	vllm/tool_parsers/granite_20b_fc_tool_parser.py	203	# re-set stuff pertaining to progress in the current tool
MEDIUM	vllm/tool_parsers/step3_tool_parser.py	51	# Explicit state flags for robust streaming
LOW	vllm/tokenizers/mistral.py	535	# if underlying tokenizer is sentencepiece, we just add "�".
LOW	vllm/platforms/cuda.py	586	# users can just use IR op priority directly
MEDIUM	vllm/distributed/utils.py	302	"""A robust barrier to synchronize all ranks.
MEDIUM	…distributed/kv_transfer/kv_connector/v1/nixl/worker.py	615	# we can leverage host_buffer for permute
MEDIUM	…ransfer/kv_connector/v1/mooncake/mooncake_connector.py	756	# Tasks can await async events, so a surplus (2x is a robust heuristic)
MEDIUM	vllm/config/model.py	310	"""Enable the custom cumem allocator to leverage advanced GPU memory
MEDIUM	vllm/config/parallel.py	580	# To make the initialization more robust we retry a few times
LOW	vllm/model_executor/layers/mamba/ops/causal_conv1d.py	185	# first chunk and does not have prior-token, so just set to 0
MEDIUM	vllm/model_executor/layers/fused_moe/modular_kernel.py	47	# The goal is to be able to utilize different communication mechanisms with
LOW	…m/model_executor/layers/fused_moe/runner/moe_runner.py	259	# Once the MK can be created upfront, we can just pass in the proper
MEDIUM	…xecutor/layers/fused_moe/prepare_finalize/deepep_ll.py	40	# TODO (varun) : Optimize leverage num_tokens_per_expert counts
MEDIUM	…executor/layers/fused_moe/experts/fused_humming_moe.py	233	# Neighboring nodes are required to utilize distinct workspaces.
MEDIUM	vllm/model_executor/layers/quantization/fp8.py	281	# For GPUs that lack FP8 hardware support, we can leverage the Marlin
MEDIUM	vllm/model_executor/layers/quantization/modelopt.py	317	# Normalize quant_algo for robust matching (ModelOpt may emit lowercase).
MEDIUM	vllm/model_executor/layers/quantization/fbgemm_fp8.py	53	# For GPUs that lack FP8 hardware support, we can leverage the Marlin
MEDIUM	…executor/layers/quantization/utils/marlin_utils_fp4.py	101	# to fully utilize the E4M3 dynamic range (e.g., global_scale=1).
MEDIUM	…executor/layers/quantization/utils/marlin_utils_fp4.py	161	# For GPUs that lack FP4 hardware support, we can leverage the
MEDIUM	…executor/layers/quantization/utils/marlin_utils_fp8.py	53	# For GPUs that lack FP8 hardware support, we can leverage the
LOW	…odel_executor/layers/quantization/utils/quant_utils.py	406	# Unquantized layer: just return base weights
MEDIUM	…/model_executor/layers/quantization/quark/quark_moe.py	166	# For GPUs that lack FP8 hardware support, we can leverage the Marlin
MEDIUM	…rs/quantization/compressed_tensors/transform/module.py	85	# do not fold into weight in order to utilize FWHT
MEDIUM	vllm/model_executor/models/grok1.py	93	# Check for Grok2-specific attributes (both for robust detection)
MEDIUM	vllm/model_executor/models/deepseek_ocr.py	133	"""Example of overriding the wrapper class `__init__()` in order to utilize
MEDIUM	vllm/model_executor/models/deepseek_ocr.py	133	"""Example of overriding the wrapper class `__init__()` in order to utilize
LOW	vllm/model_executor/models/qwen3_asr.py	464	# No audio features, just return linear positions
LOW	vllm/model_executor/models/transformers/multimodal.py	197	# NOTE: we can't just set caching=False because base class method
MEDIUM	…dels/deepseek_v4/nvidia/ops/fused_indexer_q_cutedsl.py	167	# all threads in a warp to be active since we utilize warp shuffle later.
LOW	vllm/reasoning/granite_reasoning_parser.py	198	# corrected; just return the delta text as normal content.
LOW	vllm/reasoning/hunyuan_a13b_reasoning_parser.py	91	# this id is not part of content, so just return [] here.
LOW	vllm/reasoning/olmo3_reasoning_parser.py	271	# this id is not part of content, so just return [] here.
MEDIUM	vllm/multimodal/evs.py	255	# exact timestamp count. This is robust even when early
MEDIUM	vllm/benchmarks/datasets/datasets.py	2910	# leverage CustomDataset sample
53 more matches not shown…

Verbosity Indicators147 hits · 265 pts

Severity	File	Line	Snippet
LOW	tests/v1/logits_processors/test_correctness.py	1180	# Step 1: think-start token appears.
LOW	tests/v1/kv_connector/unit/test_lmcache_connector.py	567	# Step 1: Get events from lmcache engine
LOW	tests/v1/kv_connector/unit/test_lmcache_connector.py	576	# Step 2: Update connector output (simulate receiving from worker)
LOW	tests/v1/kv_connector/unit/test_lmcache_connector.py	582	# Step 3: Take events
LOW	…/v1/kv_connector/unit/test_remote_prefill_lifecycle.py	433	# Step 2: 5 blocks are in use (2 new for remote blocks).
LOW	…/v1/kv_connector/unit/test_remote_prefill_lifecycle.py	441	# Step 3: finish recving (5 blocks in use)
LOW	…/v1/kv_connector/unit/test_remote_prefill_lifecycle.py	450	# Step 4: try to schedule, remote request is put to running list
LOW	…/v1/kv_connector/unit/test_remote_prefill_lifecycle.py	460	# Step 5: Remote request will be put back to waiting list
LOW	…/v1/kv_connector/unit/test_remote_prefill_lifecycle.py	468	# Step 6: finish the request, free it.
LOW	…/v1/kv_connector/unit/test_remote_prefill_lifecycle.py	477	# Step 7: now we can schedule (with 2 blocks computed),
LOW	…/v1/kv_connector/unit/test_remote_prefill_lifecycle.py	489	# Step 8: free everything.
LOW	…/v1/kv_connector/unit/test_remote_prefill_lifecycle.py	536	# Step 2: 3 blocks are in use,
LOW	…/v1/kv_connector/unit/test_remote_prefill_lifecycle.py	547	# Step 3: finish the request, free it.
LOW	…/v1/kv_connector/unit/test_remote_prefill_lifecycle.py	556	# Step 4: now we can initiate KV transfer (with 2 blocks computed).
LOW	…/v1/kv_connector/unit/test_remote_prefill_lifecycle.py	564	# Step 5: finish recving (5 blocks in use)
LOW	…/v1/kv_connector/unit/test_remote_prefill_lifecycle.py	573	# Step 6: schedule remote request
LOW	…/v1/kv_connector/unit/test_remote_prefill_lifecycle.py	580	# Step 7: free everything.
LOW	tests/v1/determinism/test_batch_invariance.py	700	# Step 1: Run decode and collect logprobs
LOW	tests/v1/determinism/test_batch_invariance.py	719	# Step 2: For each token position, run prefill and compare
LOW	tests/v1/streaming_input/test_scheduler_streaming.py	373	# Step 2: Schedule creates NewRequestData
LOW	tests/v1/streaming_input/test_scheduler_streaming.py	447	# Step 7: Schedule again - now request uses cached state
LOW	tests/v1/streaming_input/test_scheduler_streaming.py	514	# Step 12: Add new streaming request with seq_id=1
LOW	tests/v1/streaming_input/test_scheduler_streaming.py	388	# Step 3: Simulate model runner caching the prompt_token_ids
LOW	tests/v1/streaming_input/test_scheduler_streaming.py	421	# Step 6: Verify request state after Cycle 1
LOW	tests/v1/streaming_input/test_scheduler_streaming.py	459	# Step 8: Calculate num_tokens like gpu_model_runner.py:1284 does
LOW	tests/v1/streaming_input/test_scheduler_streaming.py	495	# Step 11: Verify request transitioned to WAITING_FOR_STREAMING_REQ
LOW	tests/v1/streaming_input/test_scheduler_streaming.py	526	# Step 13: Scheduler schedules the updated session
LOW	tests/v1/streaming_input/test_scheduler_streaming.py	544	# Step 14: Model runner caches NEW prompt_token_ids reference
LOW	tests/v1/streaming_input/test_scheduler_streaming.py	557	# Step 15: FINAL CRITICAL VERIFICATION
LOW	…/streaming_input/test_gpu_model_runner_v2_streaming.py	84	# Step 1: Add initial request with 3 prompt tokens, all computed
LOW	…/streaming_input/test_gpu_model_runner_v2_streaming.py	100	# Step 2: Create streaming update with extended prompt
LOW	…/streaming_input/test_gpu_model_runner_v2_streaming.py	116	# Step 3: Verify no free_indices leak (old slot recycled)
LOW	…/streaming_input/test_gpu_model_runner_v2_streaming.py	155	# Step 1: Add initial request with one audio feature
LOW	…/streaming_input/test_gpu_model_runner_v2_streaming.py	175	# Step 2: Create streaming update with additional multimodal feature
LOW	…/streaming_input/test_gpu_model_runner_v2_streaming.py	192	# Step 3: Verify no free_indices leak
LOW	…/v1/streaming_input/test_gpu_model_runner_streaming.py	59	# Step 1: Create initial request state with some computed tokens
LOW	…/v1/streaming_input/test_gpu_model_runner_streaming.py	77	# Step 2: Create new request data with extended prompt
LOW	…/v1/streaming_input/test_gpu_model_runner_streaming.py	96	# Step 3: Update the request
LOW	…/v1/streaming_input/test_gpu_model_runner_streaming.py	101	# Step 4: Verify the request state was updated correctly
LOW	…/v1/streaming_input/test_gpu_model_runner_streaming.py	131	# Step 1: Create initial request state with one multimodal feature
LOW	…/v1/streaming_input/test_gpu_model_runner_streaming.py	156	# Step 2: Create new request data with additional multimodal feature
LOW	…/v1/streaming_input/test_gpu_model_runner_streaming.py	176	# Step 3: Update the request
LOW	…/v1/streaming_input/test_gpu_model_runner_streaming.py	181	# Step 4: Verify the request state was updated correctly
LOW	…1/ec_connector/integration/run_epd_correctness_test.sh	459	# Step 1: Run baseline
LOW	…1/ec_connector/integration/run_epd_correctness_test.sh	462	# Step 2: Test 1E + 1PD
LOW	…1/ec_connector/integration/run_epd_correctness_test.sh	465	# Step 3: Test baseline 1P + 1D
LOW	…1/ec_connector/integration/run_epd_correctness_test.sh	468	# Step 4: Test 1E + 1P + 1D
LOW	tests/kernels/test_fused_inv_rope_fp8_quant.py	683	# Step 1: In-place CUDA RoPE (same as production)
LOW	tests/kernels/test_fused_inv_rope_fp8_quant.py	695	# Step 2: Reshape + quant + reshape (same as production)
LOW	…ultimodal/generation/test_vit_backend_functionality.py	415	# Step 1: Backend filtering
LOW	…ultimodal/generation/test_vit_backend_functionality.py	425	# Step 2: Apply GPU marks dynamically
LOW	…ultimodal/generation/test_vit_backend_functionality.py	430	# Step 3: Route to appropriate handler
LOW	…ts/entrypoints/llm/test_mm_cache_external_injection.py	91	# Step 1: Normal requests to populate the cache
LOW	…ts/entrypoints/llm/test_mm_cache_external_injection.py	99	# Step 2: Use a second image to get valid expanded tokens and
LOW	…ntrypoints/weight_transfer/test_weight_transfer_llm.py	263	# Step 1: Initialize weight transfer engine
LOW	…ntrypoints/weight_transfer/test_weight_transfer_llm.py	268	# Step 2: Start weight update
LOW	…ntrypoints/weight_transfer/test_weight_transfer_llm.py	271	# Step 3: Update weights
LOW	…ntrypoints/weight_transfer/test_weight_transfer_llm.py	282	# Step 4: Finish weight update
LOW	tests/entrypoints/openai/responses/test_harmony.py	1069	# Step 1: Get a function call from the model
LOW	tests/entrypoints/openai/responses/test_harmony.py	1095	# Step 2: Build full conversation history
87 more matches not shown…

Cross-Language Confusion50 hits · 246 pts

Severity	File	Line	Snippet
HIGH	setup.py	656	"build_tag": null,
HIGH	setup.py	660	"variant": null,
HIGH	csrc/cpu/generate_cpu_attn_dispatch.py	156	(__riscv_v_min_vlen == 128 \|\| __riscv_v_min_vlen == 256)
HIGH	csrc/cpu/generate_cpu_attn_dispatch.py	228	"&& (__riscv_v_min_vlen == 128 \|\| __riscv_v_min_vlen == 256)",
HIGH	csrc/quantization/machete/generate.py	512	"M > 256 && K <= 16384 && N <= 4096": ((128, 128), (2, 1, 1)),
HIGH	csrc/quantization/machete/generate.py	515	"M > 128 && K <= 4096 && N <= 4096": ((128, 64), (2, 1, 1)),
HIGH	csrc/quantization/machete/generate.py	516	"M > 128 && K <= 8192 && N <= 8192": ((128, 128), (2, 1, 1)),
HIGH	csrc/quantization/machete/generate.py	519	"M > 64 && K <= 4069 && N <= 4069": ((128, 32), (2, 1, 1)),
HIGH	csrc/quantization/machete/generate.py	520	"M > 64 && K <= 4069 && N <= 8192": ((128, 64), (2, 1, 1)),
HIGH	csrc/quantization/machete/generate.py	521	"M > 64 && K >= 8192 && N >= 12288": ((256, 128), (2, 1, 1)),
HIGH	csrc/quantization/machete/generate.py	524	"M > 32 && K <= 6144 && N <= 6144": ((128, 16), (1, 1, 1)),
HIGH	csrc/quantization/machete/generate.py	525	"M > 32 && K >= 16384 && N >= 12288": ((256, 64), (2, 1, 1)),
HIGH	csrc/quantization/machete/generate.py	528	"M > 16 && K <= 12288 && N <= 8192": ((128, 32), (2, 1, 1)),
HIGH	tests/v1/attention/test_trtllm_attention_integration.py	140	# Randomly permute blocks (starting from block 1; block 0 is null).
HIGH	tests/v1/attention/test_mla_backends.py	256	# Permute the context blocks (excluding block 0 which is null)
HIGH	tests/v1/attention/test_attention_backends.py	164	# Permute the context blocks (excluding block 0 which is null)
HIGH	tests/v1/core/test_scheduler.py	2630	num_blocks=5, # Can hold 64 tokens (first block is null)
HIGH	tests/v1/core/test_scheduler.py	3412	num_blocks=15, # can hold 244 tokens with 14 blocks (first block is null)
HIGH	tests/v1/core/test_scheduler.py	3653	num_blocks=11, # Can hold 160 tokens (first block is null)
HIGH	tests/v1/cudagraph/test_cudagraph_mode.py	67	# when above code raises, `llm` may be undefined, so we need to catch that
HIGH	tests/v1/cudagraph/test_cudagraph_mode.py	123	# when above code raises, `llm` may be undefined, so we need to catch that
HIGH	tests/tool_parsers/test_granite_tool_parser.py	36	"null_field": null,
HIGH	tests/tool_parsers/test_gemma4_tool_parser.py	91	# instead of `{"param": null}` for nullable tool parameters.
HIGH	tests/tool_parsers/test_gemma4_tool_parser.py	94	assert json.dumps(result) == '{"param": null}'
HIGH	tests/tool_parsers/test_granite_20b_fc_tool_parser.py	36	"null_field": null,
HIGH	tests/tool_parsers/test_olmo3_tool_parser.py	34	"role=null, "
HIGH	tests/tool_parsers/test_olmo3_tool_parser.py	43	'"role": null, '
HIGH	tests/tool_parsers/test_pythonic_tool_parser.py	35	'"role": null, '
HIGH	tests/tool_parsers/test_deepseekv3_tool_parser.py	43	""""bool_field": true, "null_field": null, """
HIGH	tests/tool_parsers/test_internlm2_tool_parser.py	54	"null_field": null,
HIGH	tests/tool_parsers/test_llama4_pythonic_tool_parser.py	35	'"role": null, '
HIGH	tests/tool_parsers/test_phi4mini_tool_parser.py	51	"null_field": null,
HIGH	tests/tool_parsers/test_lfm2_tool_parser.py	38	'"role": null, '
HIGH	tests/tool_parsers/test_longcat_tool_parser.py	54	"null_field": null,
HIGH	tests/tool_parsers/test_hunyuan_a13b_tool_parser.py	45	'<tool_calls>[{"name": "get_weather", "arguments": {"city": "San Francisco", "metric": "celsius"}}, {"name":
HIGH	tests/tool_parsers/test_qwen3coder_tool_parser.py	633	# Multi non-null: anyOf[string, integer, null] → first non-null is string
HIGH	tests/tool_parsers/test_qwen3coder_tool_parser.py	189	[{"question": "Pick a color", "multiSelect": false, "answer": null}]
HIGH	…entrypoints/serve/disagg/test_return_routed_experts.py	36	'{"sliding_window": null}',
HIGH	tests/entrypoints/openai/test_return_routed_experts.py	33	'{"sliding_window": null}',
HIGH	vllm/_custom_ops.py	1812	assert k_times_2 % 2 == 0, "input width must be even (gate \|\| up layout)"
HIGH	vllm/_custom_ops.py	1921	assert k_times_2 % 2 == 0, "input width must be even (gate \|\| up layout)"
HIGH	vllm/_custom_ops.py	1792	input_tensor: The input tensor with gate \|\| up layout [m_topk, k*2]
HIGH	vllm/v1/core/single_type_kv_cache_manager.py	777	# result [null] [null] ... [null] [hit block 1 (1st block contain
HIGH	vllm/v1/core/single_type_kv_cache_manager.py	338	every (non-null) block — the default for full attention.
HIGH	vllm/v1/core/single_type_kv_cache_manager.py	731	[null, null, block 3], otherwise, we return [null, null]
HIGH	vllm/v1/core/single_type_kv_cache_manager.py	736	we return 4 blocks[null, null, null, null]
HIGH	vllm/tool_parsers/utils.py	280	(null, true, false) that some models produce instead of Python
HIGH	…ibuted/kv_transfer/kv_connector/v1/flexkv_connector.py	47	cd FlexKV && bash build.sh
HIGH	…isaggregated/flexkv_connector/prefix_caching_flexkv.py	12	2. cd FlexKV && bash build.sh
HIGH	.buildkite/scripts/generate-nightly-index.py	209	"build_tag": null,

Docstring Block Structure39 hits · 195 pts

Severity	File	Line	Snippet
HIGH	vllm/v1/attention/backends/registry.py	208	Register or override a backend implementation. Args: backend: The AttentionBackendEnum member to register
HIGH	vllm/v1/attention/backends/mla/prefill/registry.py	102	Register or override an MLA prefill backend implementation. Args: backend: The MLAPrefillBackendEnum member
HIGH	vllm/v1/core/single_type_kv_cache_manager.py	718	For chunked local attention, we need to find the longest cache hit prefix of the blocks that is not lon
HIGH	vllm/v1/structured_output/utils.py	290	Check if grammar appears to use Lark syntax. Args: grammar_str: Input grammar string Returns:
HIGH	vllm/v1/structured_output/utils.py	322	Convert a Lark grammar string to EBNF format. EBNF reference: https://github.com/ggerganov/llama.cpp/blob/
HIGH	vllm/v1/worker/utils.py	264	Select a block size that is supported by all backends and is a factor of kv_manager_block_size. If kv_mana
HIGH	vllm/tool_parsers/apertus_tool_parser.py	137	Buffers incoming delta chunks to prevent fragmentation of multi-token special tags. If a chunk
HIGH	vllm/tool_parsers/apertus_tool_parser.py	181	Extracts tool calls from a completely generated model response (Non-Streaming). Args: mode
HIGH	vllm/tool_parsers/apertus_tool_parser.py	283	Handles streaming chunks Args: previous_text: The complete model text generated prior to t
HIGH	vllm/tool_parsers/apertus_tool_parser.py	494	Calculates the exact string difference to safely append new tool parameters. This ensures characters l
HIGH	…ed/kv_transfer/kv_connector/v1/lmcache_mp_connector.py	738	Get number of new tokens that can be loaded from the external KV cache beyond the num_computed_tokens.
HIGH	vllm/distributed/kv_transfer/kv_connector/v1/base.py	459	Get number of new tokens that can be loaded from the external KV cache beyond the num_computed_tokens.
HIGH	…ed/kv_transfer/kv_connector/v1/moriio/moriio_engine.py	205	Get remote allocation info for a request. Args: transfer_id:TransferId The request ID Retu
HIGH	…nector/v1/lmcache_integration/multi_process_adapter.py	198	Submit a new lookup request to LMCache if there is no ongoing request. Supports both token-based and h
HIGH	…nector/v1/lmcache_integration/multi_process_adapter.py	601	Check and get the finished store and retrieve requests. Args: finished_req_ids_from_engine
HIGH	…/kv_transfer/kv_connector/v1/p2p/tensor_memory_pool.py	107	Allocates a memory block of at least the requested size. Args: size (int): Minimum size of memory t
HIGH	…/kv_transfer/kv_connector/v1/p2p/tensor_memory_pool.py	184	Stores a CUDA tensor in pinned host memory. Args: tensor (torch.Tensor): CUDA tensor to store
HIGH	…/kv_transfer/kv_connector/v1/p2p/tensor_memory_pool.py	229	Loads a tensor from pinned host memory to the specified device. Args: addr (int): Address where ten
HIGH	vllm/distributed/weight_transfer/factory.py	82	Create a weight transfer engine instance. Args: config: Weight transfer configuration containing th
HIGH	vllm/distributed/weight_transfer/base.py	86	Construct typed init info from dict with validation. Args: init_dict: Dictionary containin
HIGH	vllm/distributed/weight_transfer/base.py	106	Construct typed update info from dict with validation. Args: update_dict: Dictionary conta
HIGH	vllm/logging_utils/formatter.py	22	Shortens a file path for logging display: - Removes leading 'vllm' folder if present.
HIGH	vllm/model_executor/kernels/linear/__init__.py	441	Choose a _KernelT that can implement the given config for the given compute capability. Attempts to choose the
HIGH	vllm/model_executor/kernels/linear/__init__.py	619	Choose an MPLinearKernel that can implement the given config for the given compute capability. Attempts to cho
HIGH	…/model_executor/layers/fused_moe/expert_map_manager.py	337	Map global expert ID to local expert ID. Args: global_id: Global expert ID (0 to global_nu
HIGH	vllm/model_executor/layers/fla/ops/chunk.py	153	Args: q (torch.Tensor): Queries of shape `[B, T, H, K]`. k (torch.Tensor):
HIGH	vllm/model_executor/layers/fla/ops/fused_recurrent.py	530	Args: q (torch.Tensor): queries of shape `[B, T, H, K]`. k (torch.Tensor):
HIGH	vllm/model_executor/models/keye_vl1_5.py	79	Return num_patches per video. Args: grid_thw: Tensor with shape [N, 3] containing temporal, height, wi
HIGH	vllm/model_executor/models/isaac.py	246	Apply pixel shuffle to a packed vision sequence without unpacking per image. Args: x (`torch.Tensor`):
HIGH	vllm/parser/parser_manager.py	37	Retrieve a registered or lazily registered Parser class. Args: name: The registered name o
HIGH	vllm/multimodal/audio.py	90	Normalize audio to the specified format. This function handles channel reduction for multi-channel audio, suppo
HIGH	vllm/benchmarks/lib/ready_checker.py	25	Wait for an endpoint to become available before starting benchmarks. Args: request_func: The async req
HIGH	vllm/entrypoints/chat_utils.py	1440	Parses a given multi-modal content part based on its type. Args: part: A dict containing the content p
HIGH	vllm/transformers_utils/gguf_utils.py	174	Extract vision config parameters from mmproj.gguf metadata. Reads vision encoder configuration from GGUF metadata f
HIGH	vllm/transformers_utils/processors/isaac.py	192	Convert normalized images into flattened ViT-style patches. Args: image (`torch.Tensor`): Tenso
HIGH	vllm/lora/resolver.py	72	Get a registered resolver instance by name. Args: resolver_name: Name of the resolver to get.
HIGH	…ications/chatbot/streamlit_openai_chatbot_webserver.py	111	Generate and stream LLM response with optional reasoning process. Args: messages (list): List of conversati
HIGH	benchmarks/benchmark_long_document_qa_throughput.py	68	Repeat each prompt in the list for a specified number of times. The order of prompts in the output list depends
HIGH	benchmarks/attention_benchmarks/batch_spec.py	74	Parse batch specification string into list of BatchRequest objects. Grammar: (<count>?) q<q_len>(k?) (s<seq_le

Hallucination Indicators9 hits · 95 pts

Severity	File	Line	Snippet
CRITICAL	rust/src/cmd/src/cli/unsupported.rs	138	/// - `vllm.entrypoints.cli.serve.ServeSubcommand.subparser_init(...)`
CRITICAL	tests/v1/e2e/general/test_mamba_prefix_cache.py	773	assert engine.llm_engine.engine_core.engine_core.scheduler.reset_prefix_cache()
CRITICAL	tests/distributed/test_torchrun_example_moe.py	68	llm.llm_engine.model_executor.driver_worker.worker.model_runner.model.parameters()
CRITICAL	tests/distributed/test_torchrun_example.py	59	llm.llm_engine.model_executor.driver_worker.worker.model_runner.model.parameters()
CRITICAL	tests/models/language/generation/test_gemma.py	19	lambda self: self.model_runner.model.language_model.model.normalizer.cpu().item() # noqa: E501
CRITICAL	tests/models/language/generation/test_gemma.py	24	lambda self: self.model_runner.model.model.normalizer.cpu().item()
CRITICAL	vllm/v1/spec_decode/llm_base_proposer.py	1311	self.model.model.embed_tokens.weight.cpu(),
CRITICAL	vllm/model_executor/layers/fla/ops/utils.py	170	triton.runtime.driver.active.utils.get_device_properties(i)[
CRITICAL	docs/training/layerwise.md	124	model = llm.llm_engine.engine_core.engine_core.model_executor.driver_worker.worker.get_model()

Fake / Example Data20 hits · 23 pts

Severity	File	Line	Snippet
LOW	tests/tool_parsers/test_kimi_k2_tool_parser.py	85	'{"to": "user@example.com", "subject": "Daily Update"}',
LOW	tests/tool_parsers/test_kimi_k2_tool_parser.py	92	{"to": "user@example.com", "subject": "Daily Update"},
LOW	tests/tool_parsers/test_mistral_tool_parser.py	303	"name": "John Doe",
LOW	tests/tool_parsers/test_mistral_tool_parser.py	693	"name": "John Doe",
LOW	tests/tool_parsers/test_mistral_tool_parser.py	1056	"name": "John Doe",
LOW	tests/tool_parsers/test_mistral_tool_parser.py	296	"""[TOOL_CALLS] [{"arguments":{"name": "John Doe"}, "name": "get_age"}]""", # noqa: E501
LOW	tests/tool_parsers/test_mistral_tool_parser.py	686	"""[TOOL_CALLS] [{"arguments": {"name": "John Doe"}, "name": "get_age"}]""", # noqa: E501
LOW	tests/tool_parsers/test_mistral_tool_parser.py	1049	"""[TOOL_CALLS] [{"arguments": {"name": "John Doe"}, "name": "get_age"}]""", # noqa: E501
LOW	tests/tool_parsers/test_olmo3_tool_parser.py	23	"register_user(name='John Doe', "
LOW	tests/tool_parsers/test_olmo3_tool_parser.py	31	"register_user(name='John Doe', "
LOW	tests/tool_parsers/test_olmo3_tool_parser.py	40	arguments='{"name": "John Doe", '
LOW	tests/tool_parsers/test_pythonic_tool_parser.py	23	"register_user(name='John Doe', "
LOW	tests/tool_parsers/test_pythonic_tool_parser.py	32	arguments='{"name": "John Doe", '
LOW	tests/tool_parsers/test_lfm2_tool_parser.py	26	"register_user(name='John Doe', "
LOW	tests/tool_parsers/test_lfm2_tool_parser.py	35	arguments='{"name": "John Doe", '
LOW	tests/tool_parsers/test_lfm2_tool_parser.py	339	"deliveryAddress='123 Main St')]"
LOW	tests/tool_parsers/test_hunyuan_a13b_tool_parser.py	45	'<tool_calls>[{"name": "get_weather", "arguments": {"city": "San Francisco", "metric": "celsius"}}, {"name":
LOW	tests/tool_parsers/test_hunyuan_a13b_tool_parser.py	53	"name": "John Doe",
LOW	tests/benchmarks/test_txt_slices_dataset.py	20	Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor
LOW	tests/benchmarks/test_txt_slices_dataset.py	20	Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor

Example Usage Blocks14 hits · 20 pts

Severity	File	Line	Snippet
LOW	tools/setup_deepgemm_pythons.sh	6	# Usage:
LOW	tools/vllm-rocm/generate-rocm-wheels-root-index.sh	11	# Usage:
LOW	docker/docker-bake.hcl	5	# Usage:
LOW	docker/entrypoints/test_vllm_nonroot_entrypoint.sh	9	# Usage:
LOW	…/nixl_integration/run_multi_connector_accuracy_test.sh	15	# Usage:
LOW	…nector/nixl_integration/spec_decode_acceptance_test.sh	13	# Usage:
LOW	…nixl_integration/run_multi_connector_edge_case_test.sh	15	# Usage:
LOW	examples/tool_calling/chat_with_tools_offline.py	44	# Usage:
LOW	examples/ray_serving/run_cluster.sh	8	# Usage:
LOW	examples/ray_serving/multi-node-serving.sh	11	# Example usage:
LOW	examples/generate/multimodal/mistral-small_offline.py	51	# Usage:
LOW	…nchmarks/attention_benchmarks/configs/mla_prefill.yaml	14	# Usage:
LOW	.buildkite/scripts/cache-rocm-base-wheels.sh	10	# Usage:
LOW	.buildkite/scripts/tool_call/run-bfcl-eval.sh	5	# Usage:

Magic Placeholder Names5 hits · 20 pts

Severity	File	Line	Snippet
HIGH	…les/pooling/embed/openai_embedding_long_text/client.py	22	--api-key your-api-key
HIGH	…les/pooling/embed/openai_embedding_long_text/client.py	32	--api-key your-api-key
HIGH	…les/pooling/embed/openai_embedding_long_text/client.py	44	API_KEY = "your-api-key" # Replace with your actual API key
HIGH	…es/pooling/embed/openai_embedding_long_text/service.sh	19	API_KEY=${API_KEY:-"your-api-key"}
HIGH	.github/ISSUE_TEMPLATE/400-bug-report.yml	20	Consider redacting or replacing sensitive values with placeholders like `<YOUR_TOKEN_HERE>` when sharing configura

Slop Phrases7 hits · 16 pts

Severity	File	Line	Snippet
MEDIUM	tests/models/multimodal/generation/test_common.py	64	# model arch happens to be a substring of another one, you can add a
MEDIUM	tests/models/multimodal/generation/test_common.py	85	# NOTE you can add --collect-only to any of the above commands to see
LOW	…distributed/kv_transfer/kv_connector/v1/nixl/worker.py	1974	# while processing the next batch, we make sure to only set an
MEDIUM	vllm/model_executor/models/interfaces.py	224	as a language model component.
MEDIUM	…ache/disagg_prefill_lmcache_v1/disagg_vllm_launcher.sh	24	# secure random value. This is set to a fixed value for demonstration purposes only.
MEDIUM	examples/rl/rlhf_ipc.py	17	for demonstration purposes we simply zero out the weights.
MEDIUM	.buildkite/test-amd.yaml	1	# In this file, you can add more tests to run either by adding a new step or

Synthetic Comment Markers1 hit · 8 pts

Severity	File	Line	Snippet
HIGH	…tic_prefix_caching/automatic_prefix_caching_offline.py	26	# A prompt containing a large markdown table. The table is randomly generated by GPT-4.

Overly Generic Function Names4 hits · 4 pts

Severity	File	Line	Snippet
LOW	…gated/disaggregated_serving/moriio_toy_proxy_server.py	227	async def handle_request(api: str, request: Request):
LOW	…aggregated/p2p_nccl_xpyd/disagg_proxy_p2p_nccl_xpyd.py	125	async def handle_request():
LOW	…marks/disagg_benchmarks/disagg_prefill_proxy_server.py	243	async def handle_request():
LOW	benchmarks/disagg_benchmarks/round_robin_proxy.py	16	async def handle_request(self, request):

Dead Code1 hit · 2 pts

Severity	File	Line	Snippet
MEDIUM	tests/v1/e2e/general/test_streaming_input.py	502

Score History

Severity Breakdown

Pattern Findings