Skip to main content

ov-debug

Debug OpenVINO CPU/GPU plugin inference, accuracy, performance, memory, and device compilation issues, or inspect graph transformations, using debug capabilities such as tensor dumps, execution graphs, profiling, and transformation tracing. Do NOT use for frontend model conversion failures (unsupported operators, translation errors, or source-model loading), build/configuration failures, or generic test/CI failures.

소스 정보

저장소
openvinotoolkit/openvino
최근 소스 활동
2026년 10월 2일 15:20
감지된 SKILL.md 언어
영어
스타
11,007
포크
3,458

설치 방법

기본적으로 소스를 먼저 확인하는 Prompt가 선택됩니다. 직접 명령으로 전환하거나 로컬 사본을 다운로드할 수도 있습니다.

소스 파일 검토

설치 여부를 결정하기 전에 SKILL.md와 SkillsMP에 표시된 보조 파일을 읽어 보세요.

파일 탐색기
4 개 파일

SKILL.md 표시 중

SKILL.md
소스 지침 · 읽기 전용 미리보기
name
ov-debug
description
Debug OpenVINO CPU/GPU plugin inference, accuracy, performance, memory, and device compilation issues, or inspect graph transformations, using debug capabilities such as tensor dumps, execution graphs, profiling, and transformation tracing. Do NOT use for frontend model conversion failures (unsupported operators, translation errors, or source-model loading), build/configuration failures, or generic test/CI failures.
# Debug Skill ## Scope Use this skill when CPU/GPU plugin or graph transformation debug capabilities apply to the observed symptom. Frontend conversion failures are outside its scope; use an applicable frontend skill or investigate the frontend directly. Device compilation here means compiling an OpenVINO model for CPU/GPU execution, not converting a source model into OpenVINO IR. ## Prerequisites Build flags that enable debug capabilities (check CMakeCache.txt in the build dir): - `-DENABLE_DEBUG_CAPS=ON` — CPU/GPU plugin debug env vars, transformation matcher logging ## Components | Component | Reference file to read | Routing hints | |---------------------------|---------------------------------------|-------------------------------------------------------------------------------| | openvino_intel_cpu_plugin | [@components/debug-intel-cpu-plugin.md](components/debug-intel-cpu-plugin.md) | CPU: inference issues, wrong results, slow inference, tensor dumps, execution graphs | | openvino_intel_gpu_plugin | [@components/debug-intel-gpu-plugin.md](components/debug-intel-gpu-plugin.md) | GPU: inference issues, wrong results, slow inference, tensor dumps, execution graphs | | transformations | [@components/debug-transformations.md](components/debug-transformations.md) | transformation not applied, pass not firing, transformation tracing, slow compilation, graph inspection | ## Steps 1. Match the user's symptom to the routing hints above to identify the component(s) 2. Load the component's reference file 3. Follow the instructions and recommendations for debugging
GitHub에서 보기