skills.rest
amd
amdOFFICIAL
Explore all skills by this author →
WebsiteGitHub

quark-torch-debug

Diagnose failed AMD Quark installation, quantization, and export workflows from error messages.

154|27|Updated Jun 19, 2024
Search Tags:#debugging#pytorch#quantization#cuda#onnx#rocm#quark
One-click install
npx skills add https://github.com/amd/Quark --skill quark-torch-debug
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: quark-torch-debug
Source: https://github.com/amd/Quark/tree/main/.claude/skills-impl/l1-atomic/torch/quark-torch-debug
Command: npx skills add https://github.com/amd/Quark --skill quark-torch-debug

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill helps troubleshoot failed AMD Quark workflows by turning errors, stack traces, and environment details into a clear root-cause diagnosis and the smallest safe fix.

Core Features & Use Cases

  • Installation diagnosis: Identify broken installs, wrong Python environments, missing dependencies, and version mismatches.
  • Runtime debugging: Trace quantization failures, CUDA or ROCm memory issues, model loading errors, and unexpected outputs.
  • Export and compatibility checks: Narrow down ONNX, GGUF, and transformers compatibility problems before recommending a change.
  • Use case: A user pastes a Quark traceback from a failed PTQ run, and the Skill explains the likely cause, evidence to confirm it, and the safest recovery path.

Quick Start

Paste the full Quark error message, the command you ran, and any environment details so the Skill can produce a structured diagnostic report.

Frequently Asked Questions about quark-torch-debug

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I fix AMD Quark quantization failures from PyTorch stack traces?▼

To fix AMD Quark quantization failures, paste the full PyTorch error message and environment details to generate a structured diagnostic report that classifies the failure and identifies the root cause for the smallest safe recovery path.

Why does my Quark model loading fail with a transformers compatibility error?▼

Quark model loading fails due to transformers version mismatches or missing dependencies. Diagnose the exact loading error by providing the stack trace to narrow down the specific compatibility problem before recommending a safe change.

How do I troubleshoot CUDA or ROCm memory issues during ONNX export?▼

Troubleshoot CUDA or ROCm memory issues during ONNX export by tracing the runtime failure from the stack trace. The diagnosis applies to quantization scenarios to check export compatibility and identify memory constraints.

What is the best way to diagnose broken AMD Quark installations?▼

The best way to diagnose broken AMD Quark installations is to analyze environment evidence and error messages. This identifies wrong Python environments, missing dependencies, and version mismatches to determine the smallest safe recovery path.

Can I diagnose unexpected quantization outputs for PyTorch models using error messages?▼

You can diagnose unexpected PyTorch quantization outputs by analyzing the full error messages and command used. The diagnostic process requires environment evidence to trace runtime failures and provide a structured root-cause analysis.

Related Skills

Explore More Software Engineering Skills→
amdamd
154

quark-onnx-debug

Diagnose Quark ONNX installation, calibration, quantization, and custom-op failures.

Official
Advanced
amdamd
154

quark-torch-router

Route AMD Quark user intents to the correct downstream skill or workflow.

Official
Advanced
JNZader-VaultJNZader-Vault

onnx-inference

Deploys ML models across Python, Rust, and Go using ONNX Runtime with quantization and graph transformations for edge scenarios.

Official
Advanced
amdamd
154

quark-env-preflight

Collect OS, Python, GPU, and accelerator facts for Quark installation readiness.

Official
Intermediate