Home / InternLM / lmdeploy
AI Slop Code RIsk Monitor

InternLM/lmdeploy

LMDeploy is a toolkit for compressing, deploying, and serving LLMs.

17.6 Elevated pattern signal View on GitHub

Maintenance and Governance Review

This report summarizes scanner evidence for InternLM/lmdeploy, a Python project with 8,103 GitHub stars. SynthScan examined 340,520 lines of code across 1634 source files, recording 4605 pattern matches distributed across 23 syntactic categories. The legacy adjusted pattern signal is 17.6 (Elevated pattern signal); it is not an authorship estimate.

Use the findings to ask whether maintainers can explain intent, verify dependencies, simplify unnecessary scaffolding, and focus review effort. Risk Payload v1 adds ownership, supply-chain, over-engineering, and review-effort vectors when available. These are prompts for human verification, not probabilities, security findings, or measured maintenance costs.

17.6
Adjusted Signal
17.6
Raw Pattern Score
100%
Time Factor
2026-10-09
Last Push
8.1K
Stars
Python
Language
340.5K
Lines of Code
1.6K
Files
4.6K
Pattern Hits
2026-10-12
Scan Date
0.11
HC Hit Rate

Risk Review · v3

Use each vector to guide manual verification of the underlying findings.

Cognitive Ownership Risk 36.5/100 · MEDIUM

Can maintainers explain intent, assumptions, and trade-offs? 3162 finding(s).

Hallucinated Dependency & Supply Chain Risk 12.2/100 · LOW

Are imports declared and package/API names verified? 667 finding(s).

Agent Over-Engineering Ratio 49.5/100 · MEDIUM

Are wrappers, guards, and repeated structures necessary? 3506 finding(s).

PR Review Velocity Impact 59.6/100 · HIGH

Which changes need focused human review? 4605 finding(s).

What These Metrics Mean

Adjusted Signal
Legacy pattern score normalized per 1,000 scanned lines and multiplied by the repository-level time factor. It is not an estimate of AI authorship.
Raw Pattern Score
Sum of severity- and context-weighted pattern findings before applying the time factor.
Time Factor
A coarse multiplier based on repository last-push metadata. It does not identify when individual code was written.
Pattern Hits
Count of configured pattern findings. Review each location and surrounding code before acting.
HC Hit Rate
High- and critical-severity findings per scanned file. Severity is a configured detector weight, not proof of a vulnerability or authorship.
Lines / Files
Scan coverage denominators used by the legacy normalized score.

Pattern-Signal History

This chart tracks the legacy adjusted pattern score across scan runs. Changes can reflect source changes, scan coverage, or detector/scoring versions; they do not measure AI adoption or prove that a team refactored code. Compare runs only when their scoring versions are compatible. The dashed line tracks the raw pattern-hit count.

Severity Breakdown

Groups findings by configured severity weight. Severity is not a confidence estimate, security rating, or proof that a package is missing. Verify imports against the project's manifests and inspect each finding in its local design context.

CRITICAL 0HIGH 175MEDIUM 349LOW 4081

Directory Score Breakdown

This horizontal bar chart shows where configured pattern scores accumulate across directories. Treat the bars as locations to inspect, not as evidence of authorship or code quality. Use the file-level findings to decide whether a directory needs a focused ownership, dependency, complexity, or review check.

Pattern Findings

The scanner identified 4605 distinct pattern matches across 23 syntactic categories. Each entry below represents a discrete location in the source code where the engine recorded a detector match to inspect in context. Expand any category row to inspect the individual file paths, line numbers, code snippets, and the lexical context (CODE, COMMENT, or STRING) in which each match was detected.

Reading the findings table: The Severity column indicates the configured finding weight (CRITICAL / HIGH / MEDIUM / LOW). The Context column identifies whether the match occurred inside executable code, an inline comment, or a string literal — legacy comment-context matches receive a ×1.5 score weight. The ⚡ bolt icon marks clustered matches: three or more patterns within a 10-line window, each receiving an additional ×1.5 density multiplier under the legacy scoring model.

Hyper-Verbose Identifiers2594 hits · 2502 pts
SeverityFileLineSnippetContext
LOWbenchmark/benchmark_chat_completion.py175def _dataset_longest_prefix_match(dataset: str, selected: Sequence[str]) -> str | None:CODE
LOWbenchmark/benchmark_chat_completion.py360def _cached_tokens_from_usage(usage: dict[str, Any] | None) -> int:CODE
LOWbenchmark/benchmark_serving.py35def _parse_prometheus_metrics(text: str) -> dict[tuple[str, tuple[tuple[str, str], ...]], float]:CODE
LOWbenchmark/benchmark_serving.py58def _scrape_prometheus_metrics(server_ip: str,CODE
LOWbenchmark/benchmark_serving.py93def _build_specdecode_summary(before: dict[tuple[str, tuple[tuple[str, str], ...]], float],CODE
LOWbenchmark/benchmark_serving.py159def _append_specdecode_summary_csv(model_path: str,CODE
LOWbenchmark/profile_restful_api.py159async def async_request_openai_completions(CODE
LOWbenchmark/profile_restful_api.py237async def async_request_openai_chat_completions(CODE
LOWbenchmark/profile_restful_api.py345async def async_request_sglang_generate(CODE
LOWbenchmark/profile_restful_api.py866 def _gen_random_image_data_uri(width: int = width, height: int = height) -> tuple[Image.Image, str, int]:CODE
LOW⚡lmdeploy/deepseek_v4_encoding.py114def tool_calls_from_openai_format(tool_calls):CODE
LOWlmdeploy/deepseek_v4_encoding.py125def tool_calls_to_openai_format(tool_calls):CODE
LOWlmdeploy/deepseek_v4_encoding.py469def sort_tool_results_by_call_order(messages: list[dict[str, Any]]) -> list[dict[str, Any]]:CODE
LOWlmdeploy/deepseek_v4_encoding.py694def parse_message_from_completion_text(text: str, thinking_mode: str) -> dict[str, Any]:CODE
LOWlmdeploy/_guided_decoding.py32def _json_schema_from_response_format(response_format: dict[str, Any]) -> str:CODE
LOWlmdeploy/_guided_decoding.py82def ensure_response_format_compilable(response_format: dict[str, Any]) -> None:CODE
LOWlmdeploy/tokenizer.py70 def _check_transformers_version(self, model_dir: str, trust_remote_code: bool = False):CODE
LOWlmdeploy/tokenizer.py234 def _convert_tokens_to_string_with_added_encoders(CODE
LOWlmdeploy/utils.py648def init_custom_process_group(backend=None,CODE
LOWlmdeploy/messages.py219 def convert_stop_bad_words_to_ids(self, tokenizer: Tokenizer):CODE
LOW⚡lmdeploy/deepseek_v32_encoding.py73def tool_calls_from_openai_format(tool_calls):CODE
LOW⚡lmdeploy/deepseek_v32_encoding.py82def tool_calls_to_openai_format(tool_calls):CODE
LOWlmdeploy/deepseek_v32_encoding.py357def parse_message_from_completion_text(text: str, thinking_mode: str):CODE
LOWlmdeploy/metrics/metrics_processor.py101 def increase_succeeded_requests(self):CODE
LOWlmdeploy/metrics/metrics_processor.py114 def increase_api_routed_requests(self):CODE
LOWlmdeploy/metrics/metrics_processor.py118 def decrease_api_routed_requests(self):CODE
LOWlmdeploy/metrics/stats.py90 def update_from_schedule_metrics(self, scheduled_metrics: ScheduleMetrics):CODE
LOWlmdeploy/lite/apis/calibrate.py184def make_compatible_internvl_config(model_path):CODE
LOWlmdeploy/lite/apis/calibrate.py194 def _get_non_default_generation_parameters(self):CODE
LOWlmdeploy/lite/utils/calib_dataloader.py250def get_neuralmagic_calibration(dataset, tokenizer, nsamples, seed, seqlen):CODE
LOWlmdeploy/lite/utils/memory_efficient.py45def find_modules_by_return_value(model: nn.Module, value: str) -> list[nn.Module]:CODE
LOWlmdeploy/lite/utils/memory_efficient.py197def memory_efficient_inference(model: nn.Module, offload: bool = True, device: str = 'cuda') -> None:CODE
LOWlmdeploy/lite/utils/cal_qparams.py20def cal_qparams_per_channel_absmax(w: torch.Tensor, n_bits: int, return_stats: bool = False) -> QParams:CODE
LOWlmdeploy/lite/utils/cal_qparams.py36def cal_qparams_per_channel_minmax(w: torch.Tensor, n_bits: int, return_stats: bool = False) -> QParams:CODE
LOWlmdeploy/lite/utils/cal_qparams.py58def cal_qparams_per_group_absmax(w: torch.Tensor, n_bits: int, group_size: int, return_stats: bool = False) -> QParams:CODE
LOWlmdeploy/lite/utils/cal_qparams.py79def cal_qparams_per_group_minmax(w: torch.Tensor, n_bits: int, group_size: int, return_stats: bool = False) -> QParams:CODE
LOWlmdeploy/lite/utils/cal_qparams.py105def cal_qparams_per_tensor_minmax(w: torch.Tensor, n_bits: int, return_stats: bool = False) -> QParams:CODE
LOWlmdeploy/lite/utils/cal_qparams.py125def cal_qparams_per_tensor_absmax(w: torch.Tensor, n_bits: int, return_stats: bool = False) -> QParams:CODE
LOWlmdeploy/lite/utils/batch_split.py7def split_decoder_layer_inputs(batch_size, *args: torch.Tensor | Any,CODE
LOWlmdeploy/lite/utils/batch_split.py64def concat_decoder_layer_outputs(batch_outputs: list[Any]) -> Any:CODE
LOWlmdeploy/lite/quantization/calibration.py407 def _wrap_decoder_layers_for_search(self):CODE
LOWlmdeploy/cli/utils.py518 def allow_terminate_by_client(parser):CODE
LOWlmdeploy/cli/utils.py615 def prefix_cache_state_budget(parser):CODE
LOWlmdeploy/cli/utils.py626 def prefix_cache_decode_state_interval(parser):CODE
LOWlmdeploy/cli/utils.py670 def piecewise_cudagraph_max_tokens(parser):CODE
LOWlmdeploy/cli/utils.py679 def cudagraph_capture_batch_sizes(parser):CODE
LOWlmdeploy/cli/utils.py827 def dllm_confidence_threshold(parser):CODE
LOWlmdeploy/cli/utils.py835 def enable_return_routed_experts(parser):CODE
LOWlmdeploy/cli/utils.py873 def distributed_executor_backend(parser):CODE
LOWlmdeploy/vl/engine.py15def _get_hf_config_mm_feature_dtype(hf_config) -> torch.dtype | None:CODE
LOWlmdeploy/vl/engine.py44def _resolve_mm_feature_dtype(hf_config, backend_config) -> torch.dtype | None:CODE
LOWlmdeploy/vl/engine.py63def _raise_exception_on_finish(task: asyncio.Task) -> None:CODE
LOWlmdeploy/vl/model/internvl.py21def find_closest_aspect_ratio(aspect_ratio, target_ratios, width, height, image_size):CODE
LOWlmdeploy/vl/model/preprocess_utils.py77def _expand_bundled_image_items(item: dict, token_id: int) -> list[dict]:CODE
LOWlmdeploy/vl/model/preprocess_utils.py107def _expand_bundled_video_items(item: dict, token_id: int) -> list[dict]:CODE
LOWlmdeploy/vl/model/preprocess_utils.py189def _expand_bundled_audio_items(item: dict, token_id: int) -> list[dict]:CODE
LOWlmdeploy/vl/model/llava.py30def _clip_vision_tower_load_model(self, **kwargs):CODE
LOWlmdeploy/vl/model/utils.py11def disable_transformers_logging():CODE
LOWlmdeploy/vl/model/deepseek.py14def check_deepseek_vl_install():CODE
LOWlmdeploy/vl/model/base.py383 def to_pytorch_with_input_ids(self, messages):CODE
2534 more matches not shown…
Decorative Section Separators304 hits · 1023 pts
SeverityFileLineSnippetContext
MEDIUM⚡lmdeploy/deepseek_v4_encoding.py97# ============================================================COMMENT
MEDIUM⚡lmdeploy/deepseek_v4_encoding.py99# ============================================================COMMENT
MEDIUMlmdeploy/deepseek_v4_encoding.py14# ============================================================COMMENT
MEDIUMlmdeploy/deepseek_v4_encoding.py16# ============================================================COMMENT
MEDIUMlmdeploy/deepseek_v4_encoding.py39# ============================================================COMMENT
MEDIUMlmdeploy/deepseek_v4_encoding.py41# ============================================================COMMENT
MEDIUMlmdeploy/deepseek_v4_encoding.py223# ============================================================COMMENT
MEDIUMlmdeploy/deepseek_v4_encoding.py225# ============================================================COMMENT
MEDIUMlmdeploy/deepseek_v4_encoding.py405# ============================================================COMMENT
MEDIUMlmdeploy/deepseek_v4_encoding.py407# ============================================================COMMENT
MEDIUMlmdeploy/deepseek_v4_encoding.py510# ============================================================COMMENT
MEDIUMlmdeploy/deepseek_v4_encoding.py512# ============================================================COMMENT
MEDIUMlmdeploy/deepseek_v4_encoding.py609# ============================================================COMMENT
MEDIUMlmdeploy/deepseek_v4_encoding.py611# ============================================================COMMENT
MEDIUMlmdeploy/turbomind/weight_format.py54# ---------------------------------------------------------------------------COMMENT
MEDIUMlmdeploy/turbomind/weight_format.py56# ---------------------------------------------------------------------------COMMENT
MEDIUMlmdeploy/turbomind/weight_format.py95# ---------------------------------------------------------------------------COMMENT
MEDIUMlmdeploy/turbomind/weight_format.py97# ---------------------------------------------------------------------------COMMENT
MEDIUMlmdeploy/turbomind/weight_format.py188# ---------------------------------------------------------------------------COMMENT
MEDIUMlmdeploy/turbomind/weight_format.py190# ---------------------------------------------------------------------------COMMENT
MEDIUMlmdeploy/turbomind/weight_format.py465# ---------------------------------------------------------------------------COMMENT
MEDIUMlmdeploy/turbomind/weight_format.py467# ---------------------------------------------------------------------------COMMENT
MEDIUMlmdeploy/turbomind/builders/_base.py22# ---------------------------------------------------------------------------COMMENT
MEDIUMlmdeploy/turbomind/builders/_base.py24# ---------------------------------------------------------------------------COMMENT
MEDIUMlmdeploy/turbomind/builders/_base.py38# ---------------------------------------------------------------------------COMMENT
MEDIUMlmdeploy/turbomind/builders/_base.py40# ---------------------------------------------------------------------------COMMENT
MEDIUMlmdeploy/turbomind/builders/_base.py57# ---------------------------------------------------------------------------COMMENT
MEDIUMlmdeploy/turbomind/builders/_base.py59# ---------------------------------------------------------------------------COMMENT
MEDIUMlmdeploy/turbomind/builders/_base.py126# ---------------------------------------------------------------------------COMMENT
MEDIUMlmdeploy/turbomind/builders/_base.py128# ---------------------------------------------------------------------------COMMENT
MEDIUMlmdeploy/turbomind/builders/_base.py234 # ------------------------------------------------------------------COMMENT
MEDIUMlmdeploy/turbomind/builders/_base.py236 # ------------------------------------------------------------------COMMENT
MEDIUMlmdeploy/turbomind/builders/_base.py250 # ------------------------------------------------------------------COMMENT
MEDIUMlmdeploy/turbomind/builders/_base.py252 # ------------------------------------------------------------------COMMENT
MEDIUMlmdeploy/turbomind/builders/_base.py266 # ------------------------------------------------------------------COMMENT
MEDIUMlmdeploy/turbomind/builders/_base.py268 # ------------------------------------------------------------------COMMENT
MEDIUMlmdeploy/turbomind/builders/_base.py394 # ------------------------------------------------------------------COMMENT
MEDIUMlmdeploy/turbomind/builders/_base.py396 # ------------------------------------------------------------------COMMENT
MEDIUMlmdeploy/turbomind/builders/_base.py407 # ------------------------------------------------------------------COMMENT
MEDIUMlmdeploy/turbomind/builders/_base.py409 # ------------------------------------------------------------------COMMENT
MEDIUMlmdeploy/turbomind/builders/_base.py463 # ------------------------------------------------------------------COMMENT
MEDIUMlmdeploy/turbomind/builders/_base.py465 # ------------------------------------------------------------------COMMENT
MEDIUMlmdeploy/turbomind/builders/attention.py17# ---------------------------------------------------------------------------COMMENT
MEDIUMlmdeploy/turbomind/builders/attention.py19# ---------------------------------------------------------------------------COMMENT
MEDIUMlmdeploy/turbomind/builders/attention.py76# ---------------------------------------------------------------------------COMMENT
MEDIUMlmdeploy/turbomind/builders/attention.py78# ---------------------------------------------------------------------------COMMENT
MEDIUM⚡lmdeploy/turbomind/builders/ffn.py23# ---------------------------------------------------------------------------COMMENT
MEDIUM⚡lmdeploy/turbomind/builders/ffn.py25# ---------------------------------------------------------------------------COMMENT
MEDIUMlmdeploy/turbomind/builders/ffn.py45# ---------------------------------------------------------------------------COMMENT
MEDIUMlmdeploy/turbomind/builders/ffn.py47# ---------------------------------------------------------------------------COMMENT
MEDIUMlmdeploy/turbomind/builders/ffn.py79# ---------------------------------------------------------------------------COMMENT
MEDIUMlmdeploy/turbomind/builders/ffn.py81# ---------------------------------------------------------------------------COMMENT
MEDIUMlmdeploy/turbomind/builders/linear.py27# ---------------------------------------------------------------------------COMMENT
MEDIUMlmdeploy/turbomind/builders/linear.py29# ---------------------------------------------------------------------------COMMENT
MEDIUMlmdeploy/turbomind/builders/linear.py60# ---------------------------------------------------------------------------COMMENT
MEDIUMlmdeploy/turbomind/builders/linear.py62# ---------------------------------------------------------------------------COMMENT
MEDIUMlmdeploy/turbomind/builders/linear.py104# ---------------------------------------------------------------------------COMMENT
MEDIUMlmdeploy/turbomind/builders/linear.py106# ---------------------------------------------------------------------------COMMENT
MEDIUMlmdeploy/turbomind/builders/linear.py219# ---------------------------------------------------------------------------COMMENT
MEDIUMlmdeploy/turbomind/builders/linear.py221# ---------------------------------------------------------------------------COMMENT
244 more matches not shown…
Unused Imports613 hits · 604 pts
SeverityFileLineSnippetContext
LOWbenchmark/benchmark_parser.py9CODE
LOWbenchmark/benchmark_parser.py19CODE
LOWbenchmark/benchmark_chat_completion.py14CODE
LOWbenchmark/benchmark_generate.py21CODE
LOWlmdeploy/__init__.py3CODE
LOWlmdeploy/__init__.py3CODE
LOWlmdeploy/__init__.py3CODE
LOWlmdeploy/__init__.py4CODE
LOWlmdeploy/__init__.py4CODE
LOWlmdeploy/__init__.py4CODE
LOWlmdeploy/__init__.py4CODE
LOWlmdeploy/__init__.py4CODE
LOWlmdeploy/__init__.py5CODE
LOWlmdeploy/__init__.py6CODE
LOWlmdeploy/__init__.py7CODE
LOWlmdeploy/__init__.py8CODE
LOWlmdeploy/__init__.py8CODE
LOWlmdeploy/_guided_decoding.py3CODE
LOWlmdeploy/api.py2CODE
LOWlmdeploy/utils.py521CODE
LOWlmdeploy/utils.py529CODE
LOWlmdeploy/utils.py480CODE
LOWlmdeploy/pipeline.py2CODE
LOWlmdeploy/lite/__init__.py2CODE
LOWlmdeploy/lite/__init__.py3CODE
LOWlmdeploy/lite/__init__.py4CODE
LOWlmdeploy/lite/utils/__init__.py3CODE
LOWlmdeploy/lite/utils/__init__.py3CODE
LOWlmdeploy/lite/utils/__init__.py4CODE
LOWlmdeploy/lite/utils/__init__.py4CODE
LOWlmdeploy/lite/utils/__init__.py4CODE
LOWlmdeploy/lite/utils/__init__.py4CODE
LOWlmdeploy/lite/utils/__init__.py4CODE
LOWlmdeploy/lite/utils/__init__.py4CODE
LOWlmdeploy/lite/utils/__init__.py4CODE
LOWlmdeploy/lite/utils/__init__.py4CODE
LOWlmdeploy/lite/utils/__init__.py14CODE
LOWlmdeploy/lite/utils/__init__.py15CODE
LOWlmdeploy/lite/utils/__init__.py15CODE
LOWlmdeploy/lite/utils/__init__.py15CODE
LOWlmdeploy/lite/utils/__init__.py16CODE
LOWlmdeploy/lite/utils/__init__.py17CODE
LOWlmdeploy/lite/quantization/__init__.py2CODE
LOWlmdeploy/lite/quantization/__init__.py2CODE
LOWlmdeploy/lite/quantization/__init__.py3CODE
LOWlmdeploy/lite/quantization/__init__.py3CODE
LOWlmdeploy/lite/quantization/__init__.py4CODE
LOWlmdeploy/lite/quantization/weight/__init__.py2CODE
LOWlmdeploy/lite/quantization/activation/__init__.py2CODE
LOWlmdeploy/lite/quantization/activation/__init__.py2CODE
LOWlmdeploy/lite/quantization/modules/__init__.py2CODE
LOWlmdeploy/lite/model/__init__.py3CODE
LOWlmdeploy/lite/model/__init__.py3CODE
LOWlmdeploy/lite/model/__init__.py7CODE
LOWlmdeploy/cli/__init__.py2CODE
LOWlmdeploy/vl/__init__.py2CODE
LOWlmdeploy/vl/__init__.py2CODE
LOWlmdeploy/vl/__init__.py2CODE
LOWlmdeploy/vl/__init__.py2CODE
LOWlmdeploy/vl/__init__.py2CODE
553 more matches not shown…
Cross-File Repetition77 hits · 385 pts
SeverityFileLineSnippetContext
HIGHlmdeploy/vl/model/molmo.py0build the vision part of a vlm model when backend is turbomind, or load the whole vlm model when `self.with_llm==true`STRING
HIGHlmdeploy/vl/model/llama4.py0build the vision part of a vlm model when backend is turbomind, or load the whole vlm model when `self.with_llm==true`STRING
HIGHlmdeploy/vl/model/llava.py0build the vision part of a vlm model when backend is turbomind, or load the whole vlm model when `self.with_llm==true`STRING
HIGHlmdeploy/vl/model/deepseek.py0build the vision part of a vlm model when backend is turbomind, or load the whole vlm model when `self.with_llm==true`STRING
HIGHlmdeploy/vl/model/llava_hf.py0build the vision part of a vlm model when backend is turbomind, or load the whole vlm model when `self.with_llm==true`STRING
HIGHlmdeploy/vl/model/llava_next.py0build the vision part of a vlm model when backend is turbomind, or load the whole vlm model when `self.with_llm==true`STRING
HIGHlmdeploy/vl/model/gemma3_vl.py0build the vision part of a vlm model when backend is turbomind, or load the whole vlm model when `self.with_llm==true`STRING
HIGHlmdeploy/vl/model/minicpmv.py0build the vision part of a vlm model when backend is turbomind, or load the whole vlm model when `self.with_llm==true`STRING
HIGHlmdeploy/vl/model/deepseek_vl2.py0build the vision part of a vlm model when backend is turbomind, or load the whole vlm model when `self.with_llm==true`STRING
HIGHlmdeploy/vl/model/molmo.py0extract image feature. only implement it when the backend is turbomind engine. args: messages(list[dict]): the outputs oSTRING
HIGHlmdeploy/vl/model/llama4.py0extract image feature. only implement it when the backend is turbomind engine. args: messages(list[dict]): the outputs oSTRING
HIGHlmdeploy/vl/model/llava.py0extract image feature. only implement it when the backend is turbomind engine. args: messages(list[dict]): the outputs oSTRING
HIGHlmdeploy/vl/model/deepseek.py0extract image feature. only implement it when the backend is turbomind engine. args: messages(list[dict]): the outputs oSTRING
HIGHlmdeploy/vl/model/llava_hf.py0extract image feature. only implement it when the backend is turbomind engine. args: messages(list[dict]): the outputs oSTRING
HIGHlmdeploy/vl/model/llava_next.py0extract image feature. only implement it when the backend is turbomind engine. args: messages(list[dict]): the outputs oSTRING
HIGHlmdeploy/vl/model/gemma3_vl.py0extract image feature. only implement it when the backend is turbomind engine. args: messages(list[dict]): the outputs oSTRING
HIGHlmdeploy/vl/model/minicpmv.py0extract image feature. only implement it when the backend is turbomind engine. args: messages(list[dict]): the outputs oSTRING
HIGHlmdeploy/vl/model/deepseek_vl2.py0extract image feature. only implement it when the backend is turbomind engine. args: messages(list[dict]): the outputs oSTRING
HIGHlmdeploy/serve/openai/protocol.py0tear down a previously initialized weights-update group.STRING
HIGHlmdeploy/pytorch/engine/engine.py0tear down a previously initialized weights-update group.STRING
HIGHlmdeploy/pytorch/engine/mp_engine/base_worker.py0tear down a previously initialized weights-update group.STRING
HIGHlmdeploy/pytorch/engine/mp_engine/base.py0tear down a previously initialized weights-update group.STRING
HIGHlmdeploy/pytorch/engine/executor/base_worker.py0tear down a previously initialized weights-update group.STRING
HIGHlmdeploy/pytorch/engine/executor/ray_executor.py0tear down a previously initialized weights-update group.STRING
HIGHlmdeploy/pytorch/engine/executor/base.py0tear down a previously initialized weights-update group.STRING
HIGHlmdeploy/serve/openai/endpoints/management.py0tear down a previously initialized weights-update group.STRING
HIGHlmdeploy/serve/openai/endpoints/management.py0receive weights through checkpoint-engine cuda ipc.STRING
HIGHlmdeploy/pytorch/engine/engine.py0receive weights through checkpoint-engine cuda ipc.STRING
HIGHlmdeploy/pytorch/engine/mp_engine/base_worker.py0receive weights through checkpoint-engine cuda ipc.STRING
HIGHlmdeploy/pytorch/engine/mp_engine/base.py0receive weights through checkpoint-engine cuda ipc.STRING
HIGHlmdeploy/pytorch/engine/executor/base_worker.py0receive weights through checkpoint-engine cuda ipc.STRING
HIGHlmdeploy/pytorch/engine/executor/ray_executor.py0receive weights through checkpoint-engine cuda ipc.STRING
HIGHlmdeploy/pytorch/engine/executor/mp_executor.py0receive weights through checkpoint-engine cuda ipc.STRING
HIGHlmdeploy/pytorch/engine/executor/uni_executor.py0receive weights through checkpoint-engine cuda ipc.STRING
HIGHlmdeploy/pytorch/engine/executor/base.py0receive weights through checkpoint-engine cuda ipc.STRING
HIGHlmdeploy/pytorch/messages.py0update token ids, old token ids will be added to history.STRING
HIGHlmdeploy/pytorch/strategies/dllm/sequence.py0update token ids, old token ids will be added to history.STRING
HIGHlmdeploy/pytorch/strategies/ar_spec/sequence.py0update token ids, old token ids will be added to history.STRING
HIGHlmdeploy/pytorch/strategies/ar/sequence.py0update token ids, old token ids will be added to history.STRING
HIGHlmdeploy/pytorch/strategies/dllm/step_inputs.py0shrink batch — keep only sequences at delta.indices.STRING
HIGHlmdeploy/pytorch/strategies/ar_spec/step_inputs.py0shrink batch — keep only sequences at delta.indices.STRING
HIGHlmdeploy/pytorch/strategies/ar/step_inputs.py0shrink batch — keep only sequences at delta.indices.STRING
HIGHlmdeploy/pytorch/strategies/base/step_inputs.py0shrink batch — keep only sequences at delta.indices.STRING
HIGHlmdeploy/pytorch/backends/cuda/causal_conv1d.py0return whether this selected implementation supports pcg.STRING
HIGHlmdeploy/pytorch/backends/cuda/nsa.py0return whether this selected implementation supports pcg.STRING
HIGHlmdeploy/pytorch/backends/cuda/attention/fa3.py0return whether this selected implementation supports pcg.STRING
HIGHlmdeploy/pytorch/backends/cuda/attention/default.py0return whether this selected implementation supports pcg.STRING
HIGHlmdeploy/pytorch/backends/cuda/attention/mla.py0return whether this selected implementation supports pcg.STRING
HIGHlmdeploy/pytorch/backends/cuda/attention/fa3.py0describe metadata required by this selected implementation.STRING
HIGHlmdeploy/pytorch/backends/cuda/attention/default.py0describe metadata required by this selected implementation.STRING
HIGHlmdeploy/pytorch/backends/cuda/attention/mla.py0describe metadata required by this selected implementation.STRING
HIGHlmdeploy/pytorch/models/deepseek_v4.py0we might change the number of layers so we can debug the model with less gpus.STRING
HIGHlmdeploy/pytorch/models/kimi_k2_language.py0we might change the number of layers so we can debug the model with less gpus.STRING
HIGHlmdeploy/pytorch/models/qwen3_5.py0we might change the number of layers so we can debug the model with less gpus.STRING
HIGHlmdeploy/pytorch/models/qwen3_5_moe.py0we might change the number of layers so we can debug the model with less gpus.STRING
HIGHlmdeploy/pytorch/models/interns2_mobius.py0we might change the number of layers so we can debug the model with less gpus.STRING
HIGHlmdeploy/pytorch/models/meta_moe.py0we might change the number of layers so we can debug the model with less gpus.STRING
HIGHlmdeploy/pytorch/models/qwen3_next.py0we might change the number of layers so we can debug the model with less gpus.STRING
HIGHlmdeploy/pytorch/models/deepseek_v2.py0we might change the number of layers so we can debug the model with less gpus.STRING
HIGH…/pytorch/paging/block_manager/default_block_manager.py0return if physical block can be allocated for given message.STRING
17 more matches not shown…
Excessive Try-Catch Wrapping287 hits · 327 pts
SeverityFileLineSnippetContext
LOWbenchmark/benchmark_chat_completion.py35except Exception: # noqa: BLE001 - tqdm is optional for CLI progress display.CODE
LOWbenchmark/benchmark_chat_completion.py531 except Exception as e: # noqa: BLE001 - record and keep consuming SSE.CODE
LOWbenchmark/benchmark_chat_completion.py537 except Exception as e: # noqa: BLE001 - benchmark should record failures and continue.CODE
LOWbenchmark/benchmark_chat_completion.py772 except Exception as e:CODE
LOWbenchmark/benchmark_pipeline.py58 except Exception as e:CODE
LOWbenchmark/benchmark_throughput.py58 except Exception as e:CODE
LOWbenchmark/benchmark_serving.py269 except Exception as e:CODE
LOWbenchmark/benchmark_serving.py341 except Exception as e:CODE
LOWbenchmark/benchmark_serving.py343 except Exception as e:CODE
LOWbenchmark/benchmark_generate.py56except Exception: # noqa: BLE001CODE
LOWbenchmark/benchmark_generate.py274 except Exception as e: # noqa: BLE001CODE
MEDIUMbenchmark/profile_restful_api.py766 print(f'Error applying chat template: {e}, fallback to <image> tag')CODE
MEDIUMbenchmark/profile_restful_api.py1134 print(f'Error running benchmark for request rate: {request_rate}')CODE
LOWbenchmark/profile_restful_api.py148 except Exception:CODE
LOWbenchmark/profile_restful_api.py227 except Exception:CODE
LOWbenchmark/profile_restful_api.py335 except Exception:CODE
LOWbenchmark/profile_restful_api.py412 except Exception:CODE
LOWbenchmark/profile_restful_api.py763 except Exception as e:CODE
LOWbenchmark/profile_restful_api.py794 except Exception:CODE
LOWbenchmark/profile_restful_api.py1200 except Exception as e:CODE
LOWbenchmark/profile_restful_api.py1263 except Exception as e:CODE
LOW⚡lmdeploy/deepseek_v4_encoding.py105 except Exception:CODE
LOWlmdeploy/deepseek_v4_encoding.py158 except Exception:CODE
LOWlmdeploy/model.py653 except Exception:CODE
LOWlmdeploy/model.py660 except Exception:CODE
LOWlmdeploy/model.py707 except Exception:CODE
LOWlmdeploy/model.py714 except Exception:CODE
LOWlmdeploy/model.py777 except Exception as e:CODE
LOWlmdeploy/model.py793 except Exception as e:CODE
LOWlmdeploy/model.py864 except Exception:CODE
LOWlmdeploy/model.py872 except Exception:CODE
LOWlmdeploy/utils.py481 except Exception as e:CODE
LOWlmdeploy/utils.py522 except Exception:CODE
LOWlmdeploy/utils.py530 except Exception:CODE
MEDIUMlmdeploy/utils.py519def is_deep_ep_installed():CODE
MEDIUMlmdeploy/utils.py527def is_deep_gemm_installed():CODE
LOWlmdeploy/pipeline.py405 except Exception:CODE
LOW⚡lmdeploy/deepseek_v32_encoding.py67 except Exception:CODE
LOWlmdeploy/metrics/metrics_processor.py81 except Exception as e:CODE
LOWlmdeploy/hf_configs/__init__.py25 except Exception as error:CODE
LOWlmdeploy/hf_configs/__init__.py41 except Exception:CODE
LOWlmdeploy/lite/apis/gptq.py43 except Exception:CODE
MEDIUMlmdeploy/cli/chat.py173 print(f'Error: {exc}')CODE
LOWlmdeploy/cli/cli.py131 except Exception:CODE
LOWlmdeploy/vl/engine.py69 except Exception as e:CODE
LOWlmdeploy/vl/model/builder.py81 except Exception as e:CODE
LOWlmdeploy/vl/model/utils.py44 except Exception:CODE
LOWlmdeploy/vl/model/base.py101 except Exception as e:CODE
LOWlmdeploy/vl/media/time_series.py42 except Exception as e:CODE
LOWlmdeploy/vl/media/connection.py53 except Exception as e:CODE
LOWlmdeploy/turbomind/linear.py146 except Exception:CODE
LOWlmdeploy/turbomind/linear.py298 except Exception:CODE
LOWlmdeploy/turbomind/turbomind.py785 except Exception as e:CODE
LOWlmdeploy/serve/parsers/_openai_harmony.py163 except Exception as e: # xgrammar may raise ValueError/ValidationErrorCODE
LOWlmdeploy/serve/parsers/_openai_harmony.py274 except Exception: # fmt dumps or message manipulation may failCODE
LOWlmdeploy/serve/core/chat_runner.py181 except Exception as err:CODE
LOWlmdeploy/serve/core/chat_runner.py245 except Exception as err:CODE
LOWlmdeploy/serve/core/health.py133 except Exception as e:CODE
LOWlmdeploy/serve/core/async_engine.py439 except Exception:CODE
LOWlmdeploy/serve/core/async_engine.py455 except Exception as e:CODE
227 more matches not shown…
Magic Placeholder Names50 hits · 265 pts
SeverityFileLineSnippetContext
HIGH⚡autotest/utils/anthropic_messages.py376 api_key=os.getenv('ANTHROPIC_API_KEY', 'YOUR_API_KEY'),CODE
HIGHautotest/utils/tool_reasoning_definitions.py953 'Authorization': 'Bearer YOUR_API_KEY',CODE
HIGHautotest/utils/tool_reasoning_definitions.py1115 'Authorization': 'Bearer YOUR_API_KEY',CODE
HIGHautotest/utils/restful_return_check.py215 client = OpenAI(api_key='YOUR_API_KEY', base_url=f'{url.rstrip("/")}/v1')CODE
HIGHdocs/zh_cn/llm/api_server.md90 api_key='YOUR_API_KEY',CODE
HIGHdocs/zh_cn/llm/api_server.md113 client = AsyncOpenAI(api_key='YOUR_API_KEY',CODE
HIGHdocs/zh_cn/llm/codellama.md149 api_key='YOUR_API_KEY',STRING
HIGHdocs/zh_cn/llm/api_server_tools.md36client = OpenAI(api_key='YOUR_API_KEY',base_url='http://0.0.0.0:23333/v1')CODE
HIGHdocs/zh_cn/llm/api_server_tools.md133client = OpenAI(api_key='YOUR_API_KEY', base_url='http://0.0.0.0:23333/v1')CODE
HIGHdocs/zh_cn/llm/api_server_tools.md207 client = OpenAI(api_key='YOUR_API_KEY',CODE
HIGHdocs/zh_cn/llm/api_server_tools.md378client = OpenAI(api_key='YOUR_API_KEY', base_url='http://0.0.0.0:23333/v1')CODE
HIGHdocs/zh_cn/llm/api_server_lora.md77 api_key='YOUR_API_KEY',CODE
HIGHdocs/zh_cn/advance/checkpoint_engine.md56 --api-keys YOUR_API_KEYCODE
HIGHdocs/zh_cn/advance/checkpoint_engine.md76 --api-keys YOUR_API_KEYCODE
HIGHdocs/zh_cn/advance/checkpoint_engine.md88curl -H 'Authorization: Bearer YOUR_API_KEY' \CODE
HIGHdocs/zh_cn/advance/checkpoint_engine.md131 headers={'Authorization': 'Bearer YOUR_API_KEY'},CODE
HIGHdocs/zh_cn/advance/checkpoint_engine.md154curl -X POST -H 'Authorization: Bearer YOUR_API_KEY' \CODE
HIGH⚡docs/zh_cn/advance/checkpoint_engine.md192curl -X POST -H 'Authorization: Bearer YOUR_API_KEY' \CODE
HIGH⚡docs/zh_cn/advance/checkpoint_engine.md197curl -X POST -H 'Authorization: Bearer YOUR_API_KEY' \CODE
HIGH⚡docs/zh_cn/advance/checkpoint_engine.md204curl -X POST -H 'Authorization: Bearer YOUR_API_KEY' \CODE
HIGHdocs/zh_cn/advance/structed_output.md99client = OpenAI(api_key='YOUR_API_KEY', base_url='http://0.0.0.0:23333/v1')CODE
HIGHdocs/zh_cn/advance/spec_decoding.md225 client = OpenAI(api_key='YOUR_API_KEY', base_url='http://0.0.0.0:24545/v1')CODE
HIGHdocs/zh_cn/get_started/get_started.md131 api_key='YOUR_API_KEY',CODE
HIGHdocs/zh_cn/get_started/get_started.md165client = OpenAI(api_key='YOUR_API_KEY', # A dummy api_key is requiredCODE
HIGHdocs/zh_cn/multi_modal/api_server_vl.md60client = OpenAI(api_key='YOUR_API_KEY', base_url='http://0.0.0.0:23333/v1')CODE
HIGHdocs/zh_cn/multi_modal/phi3.md55client = OpenAI(api_key='YOUR_API_KEY', base_url='http://0.0.0.0:23333/v1')CODE
HIGHdocs/en/llm/api_server.md71 api_key='YOUR_API_KEY',CODE
HIGHdocs/en/llm/api_server.md94 client = AsyncOpenAI(api_key='YOUR_API_KEY',CODE
HIGHdocs/en/llm/codellama.md149 api_key='YOUR_API_KEY',STRING
HIGHdocs/en/llm/api_server_tools.md36client = OpenAI(api_key='YOUR_API_KEY',base_url='http://0.0.0.0:23333/v1')CODE
HIGHdocs/en/llm/api_server_tools.md133client = OpenAI(api_key='YOUR_API_KEY', base_url='http://0.0.0.0:23333/v1')CODE
HIGHdocs/en/llm/api_server_tools.md207 client = OpenAI(api_key='YOUR_API_KEY',CODE
HIGHdocs/en/llm/api_server_tools.md378client = OpenAI(api_key='YOUR_API_KEY', base_url='http://0.0.0.0:23333/v1')CODE
HIGHdocs/en/llm/api_server_lora.md77 api_key='YOUR_API_KEY',CODE
HIGHdocs/en/advance/checkpoint_engine.md59 --api-keys YOUR_API_KEYCODE
HIGHdocs/en/advance/checkpoint_engine.md80 --api-keys YOUR_API_KEYCODE
HIGHdocs/en/advance/checkpoint_engine.md93curl -H 'Authorization: Bearer YOUR_API_KEY' \CODE
HIGHdocs/en/advance/checkpoint_engine.md138 headers={'Authorization': 'Bearer YOUR_API_KEY'},CODE
HIGHdocs/en/advance/checkpoint_engine.md163curl -X POST -H 'Authorization: Bearer YOUR_API_KEY' \CODE
HIGH⚡docs/en/advance/checkpoint_engine.md202curl -X POST -H 'Authorization: Bearer YOUR_API_KEY' \CODE
HIGH⚡docs/en/advance/checkpoint_engine.md207curl -X POST -H 'Authorization: Bearer YOUR_API_KEY' \CODE
HIGH⚡docs/en/advance/checkpoint_engine.md214curl -X POST -H 'Authorization: Bearer YOUR_API_KEY' \CODE
HIGHdocs/en/advance/structed_output.md99client = OpenAI(api_key='YOUR_API_KEY', base_url='http://0.0.0.0:23333/v1')CODE
HIGHdocs/en/advance/spec_decoding.md228 client = OpenAI(api_key='YOUR_API_KEY', base_url='http://0.0.0.0:24545/v1')CODE
HIGHdocs/en/get_started/get_started.md134 api_key='YOUR_API_KEY',CODE
HIGHdocs/en/get_started/get_started.md169client = OpenAI(api_key='YOUR_API_KEY', base_url='http://0.0.0.0:23333/v1')CODE
HIGHdocs/en/multi_modal/api_server_vl.md83client = OpenAI(api_key='YOUR_API_KEY', base_url='http://0.0.0.0:23333/v1')CODE
HIGHdocs/en/multi_modal/phi3.md56client = OpenAI(api_key='YOUR_API_KEY', base_url='http://0.0.0.0:23333/v1')CODE
HIGHeval/config.py57 key='YOUR_API_KEY',CODE
HIGH⚡eval/eval.py121 client = OpenAI(api_key='YOUR_API_KEY', base_url=f'{server}/v1')CODE
Deep Nesting259 hits · 242 pts
SeverityFileLineSnippetContext
LOWsetup.py63CODE
LOWsetup.py79CODE
LOWsetup.py120CODE
LOWbenchmark/benchmark_chat_completion.py266CODE
LOWbenchmark/benchmark_chat_completion.py448CODE
LOWbenchmark/benchmark_chat_completion.py1005CODE
LOWbenchmark/benchmark_guided.py527CODE
LOWbenchmark/benchmark_serving.py298CODE
LOWbenchmark/benchmark_generate.py203CODE
LOWbenchmark/benchmark_generate.py280CODE
LOWbenchmark/profile_restful_api.py159CODE
LOWbenchmark/profile_restful_api.py237CODE
LOWbenchmark/profile_restful_api.py345CODE
LOWbenchmark/profile_restful_api.py1205CODE
LOWlmdeploy/deepseek_v4_encoding.py227CODE
LOWlmdeploy/deepseek_v4_encoding.py409CODE
LOWlmdeploy/deepseek_v4_encoding.py469CODE
LOWlmdeploy/model.py649CODE
LOWlmdeploy/model.py703CODE
LOWlmdeploy/tokenizer.py152CODE
LOWlmdeploy/utils.py404CODE
LOWlmdeploy/utils.py430CODE
LOWlmdeploy/utils.py552CODE
LOWlmdeploy/deepseek_v32_encoding.py143CODE
LOWlmdeploy/archs.py45CODE
LOWlmdeploy/archs.py84CODE
LOWlmdeploy/archs.py157CODE
LOWlmdeploy/metrics/metrics_processor.py45CODE
LOWlmdeploy/metrics/stats.py146CODE
LOWlmdeploy/lite/apis/auto_awq.py18CODE
LOWlmdeploy/lite/apis/smooth_quant.py18CODE
LOWlmdeploy/lite/apis/calibrate.py201CODE
LOWlmdeploy/lite/apis/calibrate.py233CODE
LOWlmdeploy/lite/utils/calib_dataloader.py14CODE
LOWlmdeploy/lite/utils/memory_efficient.py197CODE
LOWlmdeploy/lite/utils/batch_split.py7CODE
LOWlmdeploy/cli/entrypoint.py10CODE
LOWlmdeploy/cli/chat.py146CODE
LOWlmdeploy/cli/utils.py14CODE
LOWlmdeploy/cli/utils.py908CODE
LOWlmdeploy/vl/model/internvl.py21CODE
LOWlmdeploy/vl/model/internvl.py173CODE
LOWlmdeploy/vl/model/molmo.py129CODE
LOWlmdeploy/vl/model/molmo.py151CODE
LOWlmdeploy/vl/model/preprocess_utils.py207CODE
LOWlmdeploy/vl/model/llava.py305CODE
LOWlmdeploy/vl/model/builder.py36CODE
LOWlmdeploy/vl/model/utils.py29CODE
LOWlmdeploy/vl/model/qwen2.py47CODE
LOWlmdeploy/vl/model/base.py123CODE
LOWlmdeploy/vl/model/base.py363CODE
LOWlmdeploy/vl/model/minicpmv.py200CODE
LOWlmdeploy/vl/media/time_series.py31CODE
LOWlmdeploy/turbomind/checkpoint.py188CODE
LOWlmdeploy/turbomind/tokenizer_info.py135CODE
LOWlmdeploy/turbomind/converter.py30CODE
LOWlmdeploy/turbomind/converter.py164CODE
LOWlmdeploy/turbomind/supported_models.py39CODE
LOWlmdeploy/turbomind/turbomind.py689CODE
LOWlmdeploy/turbomind/builders/_base.py292CODE
199 more matches not shown…
Empty Error Handlers42 hits · 210 pts
SeverityFileLineSnippetContext
HIGHlmdeploy/tokenizer.py168except: # noqa: E722CODE
HIGHlmdeploy/metrics/metrics_processor.py39except asyncio.CancelledError:CODE
HIGHlmdeploy/cli/utils.py945except json.JSONDecodeError:CODE
HIGHlmdeploy/vl/model/internvl.py101except OSError:CODE
HIGHlmdeploy/vl/media/video_loader.py230except OSError:CODE
HIGHlmdeploy/vl/media/video_loader.py285except OSError:CODE
HIGHlmdeploy/vl/media/video_loader.py341except OSError:CODE
HIGHlmdeploy/turbomind/linear.py146except Exception:CODE
HIGHlmdeploy/serve/core/health.py92except asyncio.CancelledError:CODE
HIGHlmdeploy/serve/core/async_engine.py437except asyncio.CancelledError:CODE
HIGHlmdeploy/serve/managers/session_manager.py90except SafeRunException:CODE
HIGHlmdeploy/serve/utils/streaming_response.py46except (asyncio.CancelledError, GeneratorExit):CODE
HIGHlmdeploy/serve/utils/streaming_response.py54except (asyncio.CancelledError, GeneratorExit):CODE
HIGHlmdeploy/serve/utils/request_cleanup.py16except (asyncio.CancelledError, GeneratorExit):CODE
HIGHlmdeploy/serve/openai/launch_server.py81except ProcessLookupError:CODE
HIGHlmdeploy/serve/openai/launch_server.py92except ProcessLookupError:CODE
HIGHlmdeploy/serve/openai/endpoints/completions.py46except AttributeError:CODE
HIGHlmdeploy/serve/openai/chat_completions/fanout.py99except BaseException: # cleanup is best-effort after caller cancellationCODE
HIGHlmdeploy/serve/openai/chat_completions/validation.py38except AttributeError:CODE
HIGHlmdeploy/pytorch/backends/cuda/turbomind_awq_modules.py41except Exception:CODE
HIGH…deploy/pytorch/backends/cuda/graph_runner/piecewise.py663except Exception:CODE
HIGHlmdeploy/pytorch/disagg/backend/mooncake.py56except Exception:CODE
HIGHlmdeploy/pytorch/engine/mp_engine/ray_engine.py107except asyncio.CancelledError:CODE
HIGHlmdeploy/pytorch/engine/mp_engine/ray_engine.py120except asyncio.CancelledError:CODE
HIGHlmdeploy/pytorch/engine/mp_engine/ray_engine.py254except asyncio.CancelledError:CODE
HIGHlmdeploy/pytorch/engine/mp_engine/zmq_rpc.py256except asyncio.CancelledError:CODE
HIGHlmdeploy/pytorch/engine/mp_engine/zmq_rpc.py542except asyncio.CancelledError:CODE
HIGHlmdeploy/pytorch/engine/cache_engine/state.py87except TypeError:CODE
HIGHtests/pytorch/engine/test_zmq_rpc.py263except RPCServerDeadError:CODE
HIGH…totest/interface/restful/test_restful_abort_request.py58except Exception:CODE
HIGHautotest/interface/restful/test_restful_generate.py837except Exception:CODE
HIGH…ce/restful/tool_parser/test_tool_call_anthropic_sdk.py257except OSError:CODE
HIGHautotest/tools/pipeline/mllm_case.py135except Exception:CODE
HIGHautotest/tools/pipeline/mllm_case.py114except Exception:CODE
HIGHautotest/utils/ascend_multinode_utils.py27except Exception:CODE
HIGHautotest/utils/tool_reasoning_definitions.py211except OSError:CODE
HIGHautotest/utils/tool_reasoning_definitions.py991except Exception:CODE
HIGHautotest/utils/tool_reasoning_definitions.py1162except Exception:CODE
HIGHautotest/utils/run_restful_chat.py1232except requests.exceptions.RequestException:CODE
HIGHautotest/utils/run_restful_chat.py418except (json.JSONDecodeError, TypeError, ValueError):CODE
HIGHautotest/utils/pipeline_chat.py507except OSError:CODE
HIGHautotest/coverage_hook/sitecustomize.py6except Exception:CODE
AI Structural Patterns136 hits · 132 pts
SeverityFileLineSnippetContext
LOWbenchmark/profile_restful_api.py1185CODE
LOWlmdeploy/model.py117CODE
LOWlmdeploy/model.py204CODE
LOWlmdeploy/model.py726CODE
LOWlmdeploy/api.py86CODE
LOWlmdeploy/utils.py648CODE
LOWlmdeploy/lite/apis/auto_awq.py41CODE
LOWlmdeploy/lite/apis/smooth_quant.py18CODE
LOWlmdeploy/lite/apis/gptq.py11CODE
LOWlmdeploy/lite/apis/calibrate.py283CODE
LOWlmdeploy/lite/utils/memory_efficient.py124CODE
LOWlmdeploy/vl/model/base.py238CODE
LOWlmdeploy/turbomind/models/utils.py225CODE
LOWlmdeploy/serve/core/async_engine.py122CODE
LOWlmdeploy/serve/core/async_engine.py464CODE
LOWlmdeploy/serve/proxy/proxy.py889CODE
LOWlmdeploy/serve/proxy/proxy.py325CODE
LOWlmdeploy/serve/openai/api_server.py257CODE
LOWlmdeploy/pytorch/config.py514CODE
LOWlmdeploy/pytorch/model_inputs.py206CODE
LOWlmdeploy/pytorch/model_inputs.py306CODE
LOWlmdeploy/pytorch/model_inputs.py442CODE
LOWlmdeploy/pytorch/messages.py360CODE
LOWlmdeploy/pytorch/messages.py477CODE
LOWlmdeploy/pytorch/kernels/cuda/pagedattention.py777CODE
LOWlmdeploy/pytorch/kernels/cuda/v4_flatten_kv.py201CODE
LOWlmdeploy/pytorch/kernels/cuda/fused_lora.py159CODE
LOWlmdeploy/pytorch/kernels/cuda/flashattention.py473CODE
LOWlmdeploy/pytorch/kernels/cuda/utils.py46CODE
LOW…eploy/pytorch/kernels/cuda/compressed_tensors_w4a16.py479CODE
LOWlmdeploy/pytorch/kernels/cuda/gated_delta_rule.py256CODE
LOWlmdeploy/pytorch/kernels/cuda/gated_delta_rule.py575CODE
LOWlmdeploy/pytorch/kernels/cuda/w8a8_triton_kernels.py379CODE
LOWlmdeploy/pytorch/kernels/cuda/causal_conv1d.py212CODE
LOWlmdeploy/pytorch/kernels/cuda/rms_norm.py18CODE
LOWlmdeploy/pytorch/kernels/cuda/flatten_kv_cache.py397CODE
LOWlmdeploy/pytorch/kernels/cuda/moe/blocked_fp8.py160CODE
LOWlmdeploy/pytorch/kernels/cuda/moe/blocked_fp8.py396CODE
LOWlmdeploy/pytorch/kernels/cuda/moe/blocked_fp8.py820CODE
LOWlmdeploy/pytorch/kernels/cuda/moe/fused_moe.py545CODE
LOWlmdeploy/pytorch/kernels/cuda/moe/fused_moe.py1137CODE
LOWlmdeploy/pytorch/kernels/cuda/moe/w8a8.py341CODE
LOWlmdeploy/pytorch/kernels/cuda/moe/v4_fp4.py228CODE
LOWlmdeploy/pytorch/kernels/cuda/moe/v4_fp4.py425CODE
LOWlmdeploy/pytorch/kernels/cuda/moe/v4_fp4.py341CODE
LOWlmdeploy/pytorch/nn/attention.py26CODE
LOWlmdeploy/pytorch/nn/norm.py33CODE
LOWlmdeploy/pytorch/nn/rotary_embedding.py180CODE
LOWlmdeploy/pytorch/nn/linear/blocked_fp8.py21CODE
LOWlmdeploy/pytorch/nn/linear/blocked_fp8.py189CODE
LOWlmdeploy/pytorch/nn/linear/blocked_fp8.py297CODE
LOWlmdeploy/pytorch/nn/linear/__init__.py29CODE
LOWlmdeploy/pytorch/nn/linear/__init__.py139CODE
LOWlmdeploy/pytorch/nn/linear/__init__.py183CODE
LOWlmdeploy/pytorch/nn/linear/__init__.py220CODE
LOWlmdeploy/pytorch/nn/linear/__init__.py307CODE
LOWlmdeploy/pytorch/nn/linear/default.py20CODE
LOWlmdeploy/pytorch/nn/linear/base.py114CODE
LOWlmdeploy/pytorch/nn/linear/static_fp8.py23CODE
LOWlmdeploy/pytorch/nn/moe/blocked_fp8.py147CODE
76 more matches not shown…
Over-Commented Block130 hits · 130 pts
SeverityFileLineSnippetContext
LOWCMakeLists.txt1# Copyright (c) 2019-2023, NVIDIA CORPORATION. All rights reserved.COMMENT
LOWbenchmark/benchmark_decode.py61if __name__ == '__main__':COMMENT
LOWbenchmark/benchmark_decode.py81 # Decoded 15991314 tokens in 242.7 seconds, 65893.38488718234 tokens/s.COMMENT
LOWlmdeploy/utils.py441 # Check for CUDA version and device compute capability.COMMENT
LOWlmdeploy/pytorch/utils.py161 # When using `remote_code` in HF components like tokenizer or configCOMMENT
LOWlmdeploy/pytorch/kernels/cuda/v4_compressor.py1121COMMENT
LOWlmdeploy/pytorch/kernels/cuda/flashattention.py141 k1_ptrs = tl.advance(k1_ptrs, (0, BLOCK_N))COMMENT
LOWlmdeploy/pytorch/strategies/dllm/step_inputs.py1# Copyright (c) OpenMMLab. All rights reserved.COMMENT
LOWlmdeploy/pytorch/strategies/ar_spec/step_inputs.py1# Copyright (c) OpenMMLab. All rights reserved.COMMENT
LOWlmdeploy/pytorch/strategies/ar/step_inputs.py1# Copyright (c) OpenMMLab. All rights reserved.COMMENT
LOWlmdeploy/pytorch/backends/dlinfer/ascend/op_backend.py321 is_graph = cls.enable_graph and global_is_decoding and (is_decoding or is_multi_token_decoding)COMMENT
LOWlmdeploy/pytorch/models/kimi_k25_vision.py1# Copyright (c) OpenMMLab. All rights reserved.COMMENT
LOWlmdeploy/pytorch/models/kimi_k25_vision.py21# See the License for the specific language governing permissions andCOMMENT
LOWtests/test_lmdeploy/test_model.py281# @pytest.mark.parametrize('has_user_sys', [True, False])COMMENT
LOWtests/test_lmdeploy/test_model.py301# 'role': 'user',COMMENT
LOWtests/pytorch/test_internvl_lora.py21COMMENT
LOWdocs/zh_cn/conf.py1#COMMENT
LOWdocs/zh_cn/conf.py121master_doc = 'index'COMMENT
LOWdocs/zh_cn/conf.py161 'use_edit_page_button': True,COMMENT
LOWdocs/zh_cn/conf.py201#COMMENT
LOWdocs/zh_cn/conf.py221 # The font size ('10pt', '11pt' or '12pt').COMMENT
LOWdocs/en/conf.py1#COMMENT
LOWdocs/en/conf.py121master_doc = 'index'COMMENT
LOWdocs/en/conf.py161 'use_source_button': True,COMMENT
LOWdocs/en/conf.py201# The default sidebars (for documents that don't match any pattern) areCOMMENT
LOWdocs/en/conf.py221 #COMMENT
LOWsrc/CMakeLists.txt1# Copyright (c) 2019-2023, NVIDIA CORPORATION. All rights reserved.COMMENT
LOWsrc/turbomind/turbomind.h1// Copyright (c) OpenMMLab. All rights reserved.COMMENT
LOWsrc/turbomind/CMakeLists.txt1# Copyright (c) 2019-2023, NVIDIA CORPORATION. All rights reserved.COMMENT
LOWsrc/turbomind/kernels/CMakeLists.txt1# Copyright (c) 2019-2023, NVIDIA CORPORATION. All rights reserved.COMMENT
LOWsrc/turbomind/kernels/attention/impl_simt.h1// Copyright (c) OpenMMLab. All rights reserved.COMMENT
LOWsrc/turbomind/kernels/attention/iterator_sm80.h1// Copyright (c) OpenMMLab. All rights reserved.COMMENT
LOWsrc/turbomind/kernels/attention/impl_884.h41 static constexpr int V_N = HeadDim / OP_N; // 8COMMENT
LOWsrc/turbomind/kernels/core/floating_point.h141// FloatingPoint<5, 2>::max_denormal;COMMENT
LOWsrc/turbomind/kernels/core/common.h1// Copyright (c) OpenMMLab. All rights reserved.COMMENT
LOWsrc/turbomind/kernels/core/common.h21#endifCOMMENT
LOWsrc/turbomind/kernels/core/common.h41#if defined(__CUDACC_RTC__) || (defined(__clang__) && defined(__CUDA__))COMMENT
LOWsrc/turbomind/kernels/core/common.h61#define TM_HOST __forceinline__ __host__COMMENT
LOWsrc/turbomind/kernels/gemm/iterator_sm70.h1// Copyright (c) OpenMMLab. All rights reserved.COMMENT
LOWsrc/turbomind/kernels/gemm/tiled_mma.h1// Copyright (c) OpenMMLab. All rights reserved.COMMENT
LOWsrc/turbomind/kernels/gemm/kernel_impl.h1// Copyright (c) OpenMMLab. All rights reserved.COMMENT
LOWsrc/turbomind/kernels/gemm/kernel_impl_sm90_mixed.h1// Copyright (c) OpenMMLab. All rights reserved.COMMENT
LOWsrc/turbomind/kernels/gemm/moe_gate_python_bind.cpp1// src/turbomind/kernels/gemm/moe_gate_python_bind.cppCOMMENT
LOWsrc/turbomind/kernels/gemm/gmma_fp8_sm90.h1#pragma onceCOMMENT
LOWsrc/turbomind/kernels/gemm/sm90_mixed_pack.h1// Copyright (c) OpenMMLab. All rights reserved.COMMENT
LOWsrc/turbomind/kernels/gemm/sm90_mixed_pack.h21#endifCOMMENT
LOWsrc/turbomind/kernels/gemm/sm90_mixed_pack.h301COMMENT
LOWsrc/turbomind/kernels/gemm/sm90_mixed_traits.h1// Copyright (c) OpenMMLab. All rights reserved.COMMENT
LOWsrc/turbomind/kernels/gemm/gemm_universal_sm90_v3.h1#pragma onceCOMMENT
LOWsrc/turbomind/kernels/gemm/gemm_universal_sm90_v3.h21#include "cutlass/cutlass.h"COMMENT
LOWsrc/turbomind/kernels/gemm/mainloop_sm80_v2.h1// Copyright (c) OpenMMLab. All rights reserved.COMMENT
LOWsrc/turbomind/kernels/gemm/gemm_universal.h1// Copyright (c) OpenMMLab. All rights reserved.COMMENT
LOW…/kernels/gemm/gemm_universal_sm90_mxfp4_fp8_unfolded.h1// Copyright (c) OpenMMLab. All rights reserved.COMMENT
LOW…/kernels/gemm/gemm_universal_sm90_mxfp4_fp8_unfolded.h21#include "src/turbomind/kernels/core/sync.h"COMMENT
LOWsrc/turbomind/kernels/gemm/iterator_sm80.h1// Copyright (c) OpenMMLab. All rights reserved.COMMENT
LOW…urbomind/kernels/gemm/prepare_moe_tma_descs_sm90_fp8.h1#pragma onceCOMMENT
LOWsrc/turbomind/kernels/gemm/matrix_ptr.h101COMMENT
LOWsrc/turbomind/kernels/gemm/gemm_universal_sm90_bf16.h61#include <type_traits>COMMENT
LOWsrc/turbomind/kernels/gemm/gemm_universal_sm90_bf16.h81#include "src/turbomind/kernels/core/array.h"COMMENT
LOWsrc/turbomind/kernels/gemm/gemm_universal_sm90_mixed.h21 * contiguous [gate64|up64] range. Other consumers retain CuTe's nativeCOMMENT
70 more matches not shown…
Modern Structural Boilerplate62 hits · 62 pts
SeverityFileLineSnippetContext
LOWlmdeploy/__init__.py10__all__ = [CODE
LOWlmdeploy/lite/utils/__init__.py19__all__ = [CODE
LOWlmdeploy/lite/quantization/__init__.py6__all__ = ['WeightQuantizer', 'ActivationObserver', 'KVCacheObserver', 'CalibrationContext', 'CalibrationContextV2']CODE
LOWlmdeploy/lite/quantization/weight/__init__.py4__all__ = ['WeightQuantizer']CODE
LOWlmdeploy/lite/quantization/activation/__init__.py4__all__ = ['ActivationObserver', 'KVCacheObserver']CODE
LOWlmdeploy/lite/quantization/modules/__init__.py4__all__ = ['WeightOnlyQLinear']CODE
LOWlmdeploy/lite/model/__init__.py9__all__ = ['MODELS']CODE
LOWlmdeploy/cli/__init__.py4__all__ = ['run']CODE
LOWlmdeploy/vl/__init__.py13__all__ = [CODE
LOWlmdeploy/turbomind/tokenizer_info.py25logger = logging.getLogger(__name__)CODE
LOWlmdeploy/turbomind/linear.py27__all__ = ['ExecPlan', 'Linear', 'Weight', 'WeightPlan', 'get_linear']CODE
LOW⚡lmdeploy/turbomind/builders/ffn.py19__all__ = [CODE
LOWlmdeploy/turbomind/builders/__init__.py18__all__ = [CODE
LOWlmdeploy/serve/__init__.py6__all__ = [CODE
LOWlmdeploy/serve/parsers/_openai_harmony.py279 def _set_response_format(self, grammar: dict) -> None:CODE
LOWlmdeploy/serve/parsers/__init__.py6__all__ = ['ResponseParser', 'ResponseParserManager', 'GptOssResponseParser', 'validate_parser_names']CODE
LOWlmdeploy/serve/parsers/response_parser.py613 def _update_reasoning_tokens(self, token_ids: list[int]) -> None:CODE
LOWlmdeploy/serve/parsers/tool_parser/__init__.py15__all__ = [CODE
LOWlmdeploy/serve/parsers/reasoning_parser/__init__.py7__all__ = [CODE
LOWlmdeploy/serve/core/__init__.py7__all__ = [CODE
LOWlmdeploy/serve/managers/__init__.py4__all__ = ['Session', 'SessionManager']CODE
LOWlmdeploy/serve/processors/__init__.py4__all__ = ['MultimodalProcessor']CODE
LOWlmdeploy/serve/anthropic/__init__.py6__all__ = ['create_anthropic_router']CODE
LOWlmdeploy/serve/openai/endpoints/__init__.py7__all__ = ['create_openai_router']CODE
LOWlmdeploy/serve/openai/responses/__init__.py13__all__ = [CODE
LOWlmdeploy/serve/openai/chat_completions/__init__.py13__all__ = ['register']CODE
LOWlmdeploy/pytorch/prefix_cache_state.py134 def set_fresh_block_range(self, start_block_idx: int, end_block_idx: int) -> None:CODE
LOWlmdeploy/pytorch/configurations/__init__.py7__all__ = []CODE
LOWlmdeploy/pytorch/kernels/__init__.py10__all__ = [CODE
LOWlmdeploy/pytorch/kernels/cuda/__init__.py14__all__ = [CODE
LOWlmdeploy/pytorch/kernels/cuda/turbo_quant.py14logger = logging.getLogger(__name__)CODE
LOWlmdeploy/pytorch/kernels/default/__init__.py5__all__ = [CODE
LOWlmdeploy/pytorch/kernels/dlinfer/__init__.py15__all__ = [CODE
LOWlmdeploy/pytorch/kv_connector/__init__.py16__all__ = [CODE
LOWlmdeploy/pytorch/kv_connector/mooncake/store/worker.py231 def _setup_store(self, store: Any, local_hostname: str) -> None:CODE
LOW…deploy/pytorch/kv_connector/mooncake/store/protocol.py11__all__ = ['LOOKUP_MSG', 'RESP_ERR']CODE
LOWlmdeploy/pytorch/kv_connector/mooncake/store/lookup.py366__all__ = ['LookupKeyClient', 'LookupKeyServer', 'prepare_lookup_rpc_path']CODE
LOW…deploy/pytorch/kv_connector/mooncake/store/__init__.py6__all__ = [CODE
LOWlmdeploy/pytorch/backends/cp_utils.py130def update_dcp_metadata(attn_metadata, step_context) -> None:CODE
LOWlmdeploy/pytorch/backends/cuda/attention/fa3.py130def update_fa3_metadata(attn_metadata, step_context, **kwargs) -> None:CODE
LOW…eploy/pytorch/backends/cuda/comm/symm_mem_allgather.py15logger = logging.getLogger(__name__)CODE
LOWlmdeploy/pytorch/backends/cuda/graph_runner/runner.py96def _update_deepep_mode(context: StepContext) -> None:CODE
LOWlmdeploy/pytorch/backends/cuda/graph_runner/__init__.py5__all__ = ['CUDAGraphRunner', 'eager_boundary']CODE
LOWlmdeploy/pytorch/memdecode/__init__.py6__all__ = [CODE
LOWlmdeploy/pytorch/models/__init__.py4__all__ = ['QLinear', 'QRMSNorm']CODE
LOWlmdeploy/pytorch/models/kimi_k25_vision.py809__all__ = [CODE
LOWlmdeploy/pytorch/disagg/backend/__init__.py18__all__ = ['DLSlimeBackend', 'MooncakeBackend']CODE
LOWlmdeploy/pytorch/spec_decode/__init__.py66__all__ = ['build_spec_agent']CODE
LOWlmdeploy/pytorch/multimodal/__init__.py4__all__ = ['MultiModalData']CODE
LOWlmdeploy/pytorch/paging/__init__.py4__all__ = ['Scheduler']CODE
LOWlmdeploy/pytorch/paging/scheduler.py470 def update_connector_output(self, connector_output) -> None:CODE
LOWlmdeploy/pytorch/paging/block_trie/__init__.py7__all__ = ['BlockTrie', 'Node', 'PrefixCacheStats']CODE
LOWlmdeploy/pytorch/transformers/__init__.py4__all__ = ['config_from_pretrained', 'register_config']CODE
LOWlmdeploy/pytorch/engine/__init__.py5__all__ = ['Engine', 'EngineInstance']CODE
LOWlmdeploy/pytorch/engine/cache_engine/__init__.py8__all__ = [CODE
LOWlmdeploy/pytorch/engine/executor/base.py436 def update_configs(self) -> None:CODE
LOWlmdeploy/pytorch/devices/__init__.py4__all__ = ['DeviceContext', 'DefaultContext', 'get_device_manager']CODE
LOWtests/turbomind/linear_attn/cases.py179 def set_logical(self, layer: int, state: torch.Tensor) -> None:CODE
LOW…totest/interface/restful/test_restful_abort_request.py32def _set_log_hook(hook: Callable[[dict], None] | None) -> None:CODE
LOW⚡autotest/utils/tool_reasoning_definitions.py224def _setdefault_enable_thinking(extra_body: dict, enable_thinking: bool) -> None:CODE
2 more matches not shown…
Redundant / Tautological Comments14 hits · 23 pts
SeverityFileLineSnippetContext
LOWbenchmark/profile_restful_api.py307 # Check if this chunk contains contentCOMMENT
LOWbenchmark/profile_restful_api.py537 # Check if the cache file already existsCOMMENT
LOWlmdeploy/deepseek_v4_encoding.py349 # Check if previous message has a task - if so, this is a task output (no thinking)COMMENT
LOWlmdeploy/model.py781 # Verify if the model can perform apply_chat_template with different roles.COMMENT
LOW⚡lmdeploy/lite/apis/calibrate.py153 # Check if the child is an instance of the given layer typeCOMMENT
LOW⚡lmdeploy/lite/apis/calibrate.py161 # Check if the child contains the target module typeCOMMENT
LOW⚡lmdeploy/lite/apis/calibrate.py164 # Check if the child matches the head nameCOMMENT
LOWlmdeploy/lite/utils/batch_split.py102 # Check if the current element is a past key-value pair.COMMENT
LOWlmdeploy/serve/processors/multimodal.py264 # Check if multimodal input existsCOMMENT
LOW…loy/serve/openai/chat_completions/logits_processors.py30 # Check if token_id is within the vocab sizeCOMMENT
LOWlmdeploy/pytorch/engine/executor/ray_executor.py690 # Check if all elements in ips are present in worker_ips and vice versa (ignoring order)COMMENT
LOWautotest/utils/run_restful_chat.py113 # Check if process is still runningCOMMENT
LOWautotest/utils/run_restful_chat.py1260 # Check if process is still runningCOMMENT
LOWautotest/utils/ray_distributed_utils.py58 # Check if port is openCOMMENT
Self-Referential Comments11 hits · 19 pts
SeverityFileLineSnippetContext
MEDIUMlmdeploy/api.py97 """This function is deprecated and no longer available.STRING
MEDIUMlmdeploy/api.py110 """This function is deprecated and no longer available.STRING
MEDIUMlmdeploy/lite/quantization/modules/linear.py16 """This class implements weight only quantization linear.STRING
MEDIUMlmdeploy/turbomind/tokenizer_info.py3"""This module provides the tokenizer info class to handle the tokenizerSTRING
MEDIUMlmdeploy/pytorch/ray.py127 # Create a new placement groupCOMMENT
MEDIUMlmdeploy/pytorch/tools/utils.py165 # Create a clean table-like formatCOMMENT
MEDIUMlmdeploy/pytorch/models/llama4.py751 """This function is used to fetch the first embedding layer to activateSTRING
MEDIUMlmdeploy/pytorch/models/kimi_k25_vision.py3# This file contains code adapted from moonshotai/Kimi-K2.6COMMENT
MEDIUMlmdeploy/pytorch/models/gemma3_vl.py240 # Create a global causal mask.STRING
MEDIUMlmdeploy/pytorch/models/gemma3_vl.py261 # Create a local causal mask with sliding window (1024).STRING
MEDIUM…_lite/test_quantization/test_utils/test_cal_qparams.py18 # Create a dummy tensorCOMMENT
Docstring Block Structure3 hits · 15 pts
SeverityFileLineSnippetContext
HIGHlmdeploy/api.py24Create a pipeline for inference. Args: model_path: the path of a model. It could be one of the following opSTRING
HIGHlmdeploy/lite/utils/memory_efficient.py46Finds modules in model that return given value. Args: model (nn.Module): Model to inspect value (stSTRING
HIGHlmdeploy/serve/processors/multimodal.py229Process prompt and return prompt string and input_ids. Handles both text-only and multimodal prompts. MultimodaSTRING
Cross-Language Confusion3 hits · 12 pts
SeverityFileLineSnippetContext
HIGHlmdeploy/turbomind/parallel_config.py8 attn_tp == tp || mlp_tp == tpSTRING
HIGHlmdeploy/turbomind/parallel_config.py9 tp % attn_tp == 0 && tp % mlp_tp == 0STRING
HIGH⚡…test_lmdeploy/serve/parsers/test_json_value_scanner.py39 '{"nested":[{"text":"{}[]"},true,null,1.5]}',CODE
Structural Annotation Overuse7 hits · 11 pts
SeverityFileLineSnippetContext
LOWlmdeploy/deepseek_v32_encoding.py355# NOTE: This function parses only correctly formatted strings and will not attempt to correctCOMMENT
LOW⚡…rve/openai/chat_completions/test_delta_tool_call_id.py25 # Step 1: feed partial JSON with nameCOMMENT
LOW⚡…ts/pytorch/spec_decode/test_guided_spec_integration.py481 # Step 2: original should be at correct stateCOMMENT
LOWdocs/en/advance/update_weights.md8## Step 1: Launch serverCOMMENT
LOWdocs/en/advance/update_weights.md16## Step 2: Offloads weights & kv cacheCOMMENT
LOWdocs/en/advance/update_weights.md41## Step 3: Update weightsCOMMENT
LOWdocs/en/advance/update_weights.md74## Step 4: Wakeup serverCOMMENT
Dead Code4 hits · 8 pts
SeverityFileLineSnippetContext
MEDIUM…oy/serve/openai/chat_completions/test_n_completions.py331CODE
MEDIUM…oy/serve/openai/chat_completions/test_n_completions.py418CODE
MEDIUM…oy/serve/openai/chat_completions/test_n_completions.py475CODE
MEDIUMtests/pytorch/engine/test_ray_mp_engine.py166CODE
Modern AI Meta-Vocabulary2 hits · 6 pts
SeverityFileLineSnippetContext
MEDIUMautotest/tools/pipeline/mllm_case.py374 # In-context few-shot learningCOMMENT
MEDIUMdocs/en/multi_modal/vl_pipeline.md66### Set context window sizeCOMMENT
Verbosity Indicators3 hits · 4 pts
SeverityFileLineSnippetContext
LOWlmdeploy/serve/core/async_engine.py812 The purpose of this function is to allow users to choose whether to use the synchronous interface or theSTRING
LOW⚡…rve/openai/chat_completions/test_delta_tool_call_id.py25 # Step 1: feed partial JSON with nameCOMMENT
LOW⚡…ts/pytorch/spec_decode/test_guided_spec_integration.py481 # Step 2: original should be at correct stateCOMMENT
AI Slop Vocabulary1 hit · 3 pts
SeverityFileLineSnippetContext
MEDIUMlmdeploy/pytorch/kernels/cuda/moe/blocked_fp8.py645 # tokens. A compact BM64 schedule avoids that cliff and remains robustCOMMENT
Example Usage Blocks1 hit · 2 pts
SeverityFileLineSnippetContext
LOWtests/pytorch/engine/test_zmq_rpc.py94 # Example usageCOMMENT
Fake / Example Data1 hit · 1 pts
SeverityFileLineSnippetContext
LOWlmdeploy/model.py844 messages = [{'role': 'user', 'content': 'placeholder'}, {'role': 'assistant', 'content': 'sentinel'}]CODE
Overly Generic Function Names1 hit · 1 pts
SeverityFileLineSnippetContext
LOWtests/turbomind/linear/run_sm90_scan.py281def run_task(CODE