Fine-Tuning Expert
Guides an AI agent through fine-tuning LLMs with LoRA/QLoRA, dataset prep, hyperparameters, evaluation and deployment
- Repository
- Jeffallan/claude-skills
- Stars
- 12,176
- Licence
- MIT
- Files
- 6 files
What it does
This skill turns the agent into a guide for customizing existing AI language models on your own data. When you ask for help adapting a model, it follows a set workflow: check and format your training data, pick a lightweight method such as LoRA or QLoRA, set training settings, compare the result against the original model, and merge and optimize it for serving. It also supplies example Python code using Hugging Face libraries and reference notes on each stage.
Great for
- ML engineers or developers who want a structured starting point for LoRA or QLoRA fine-tuning
- Teams preparing JSONL instruction datasets and needing validation and deduplication steps
- People who want guidance on evaluating a fine-tuned model against its base model before deployment
Skip it if
- Beginners who only want to use an existing chat model; no training is needed for that and the skill assumes ML experience
- Large-scale or production training without expert review: the example code and advice are generic and need checking against your hardware, licences and data
- Work on models whose licence or terms forbid fine-tuning or redistribution of derived weights, since the skill does not check licences
Try asking
- “Help me set up a LoRA fine-tune of a 7B model on my customer-support JSONL file”
- “Validate and deduplicate my training dataset before I start training”
- “Recommend learning rate, batch size and rank for a small instruction-tuning run on one GPU”
- “Show me how to merge my LoRA adapter into the base model and benchmark its latency”
Install in 3 steps
- 1
Download the skill
The ZIP contains the 6 original files, the MIT licence and an attribution note.
fine-tuning-expert.zip - 2
Add it to your agent
Claude Code (all projects)
mkdir -p ~/.claude/skills && unzip fine-tuning-expert.zip -d ~/.claude/skills/Claude Code / Agent SDK (this project only)
mkdir -p .claude/skills && unzip fine-tuning-expert.zip -d .claude/skills/Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.
- 3
Ask for the task
No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.
Get better results
- Provide your dataset format, base model name, and GPU memory up front so the agent can choose between LoRA and QLoRA sensibly
- Ask for the validation and evaluation scripts explicitly, since the skill expects them but does not ship them
- Review every command the agent runs, especially those that download models or write files, and keep credentials such as Hugging Face tokens out of shared chats
What's inside
- references/dataset-preparation.md
- references/deployment-optimization.md
- references/evaluation-metrics.md
- references/hyperparameter-tuning.md
- references/lora-peft.md
- SKILL.md
Redistributed unmodified from Jeffallan/claude-skills under the MIT licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.