Diagnose Quark ONNX installation, calibration, quantization, and custom-op failures.
npx skills add https://github.com/amd/Quark --skill quark-onnx-debug
Please help me install this Agent Skill. Skill: quark-onnx-debug Source: https://github.com/amd/Quark/tree/main/.claude/skills-impl/l1-atomic/onnx/quark-onnx-debug Command: npx skills add https://github.com/amd/Quark --skill quark-onnx-debug
This Skill helps diagnose and resolve failures in AMD Quark's ONNX workflow, including installation issues, runtime errors, calibration crashes, custom-op compilation problems, and unexpected quantization results.
onnx, onnxruntime, and Quark dependencies.Paste the full Quark ONNX error message, stack trace, exact call, model details, and environment info so the Skill can produce a structured diagnostic report and smallest safe fix.
High-intent search queries and answers about installing and using this skill.
To fix Quark ONNX quantization and calibration failures, provide the full error trace, model details, and environment info to diagnose calibration reader bugs, OOM conditions, and unsupported settings for a structured recovery report.
ONNXRuntime silently falls back to CPU when execution providers are unavailable or misconfigured; debugging requires validating your environment details and provider setup to resolve CUDA or ROCm configuration issues.
Resolving custom op compilation errors when exporting PyTorch models to ONNX involves validating custom-op load failures, ABI mismatches, and graph or shape validation errors using your exact call stack and model context.
Yes, Quark ONNX debugging supports NPU execution providers, applying to PyTorch-to-ONNX and ONNX graph optimization scenarios across CPU, CUDA, ROCm, and NPU environments to validate versions and provider compatibility.
The best way to diagnose onnxruntime dependency conflicts during Quark installation is to analyze environment details and version info to identify missing or conflicting onnx, onnxruntime, and Quark packages.
cericoTrace execution paths to diagnose errors across code, APIs, and UI.
amdDiagnose failed AMD Quark installation, quantization, and export workflows from error messages.
amdBuild hardware-aware Quark ONNX quantization plans from model analysis and deployment intent.
amdExtracts ONNX model metadata needed for Quark PTQ planning and compatibility assessment.