train-debug

Community

Diagnose and fix ML training issues fast.

AuthorGonglitian
Version1.0.0
Installs0

System Documentation

What problem does it solve?

ML training often encounters crashes, non-converging losses, NaN/Inf gradients, OOM, and slowdowns that block progress. This skill provides a structured diagnostic workflow to identify root causes and guide efficient fixes.

Core Features & Use Cases

  • Structured triage workflow that maps symptoms to root causes and recommended actions.
  • Category-specific diagnosis with concrete fixes, verification steps, and best practices for multi-GPU training.
  • Clear guidance for producing reproducible results and validating fixes in production-like pipelines.

Quick Start

Describe the observed issue, run triage steps to identify root cause, and apply a safe, incremental fix while verifying results.

Dependency Matrix

Required Modules

None required

Components

Standard package

💻 Claude Code Installation

Recommended: Let Claude install automatically. Simply copy and paste the text below to Claude Code.

Please help me install this Skill:
Name: train-debug
Download link: https://github.com/Gonglitian/agent-skills/archive/main.zip#train-debug

Please download this .zip file, extract it, and install it in the .claude/skills/ directory.
View Source Repository

Agent Skills Search Helper

Install a tiny helper to your Agent, search and equip skill from 471,000+ vetted skills library on demand.