peft-fine-tuning

Parameter-efficient fine-tuning for LLMs using LoRA, QLoRA, and 25+ methods. Use when fine-tuning large models (7B-70B) with limited GPU memory, when you need to train <1% of parameters with minimal accuracy loss, or for multi-adapter serving. HuggingFace's official library integrated with transformers ecosystem.

Quality

81%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Passed

No findings from the security scan

Quality

Content

76%Weight 40%Scale 1-5

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

A comprehensive, highly actionable PEFT reference with executable code throughout and properly signaled reference files. The main gap is the absence of explicit validation/feedback checkpoints in the training workflow, which the rubric caps at 3.

Suggestions

Add explicit validation checkpoints to the quick-start training workflow (e.g., inspect training loss, run evaluation on held-out data, verify adapter loads before merging) framed as validate→fix→retry loops to lift workflow_clarity above the batch-operation cap of 3.

Move the 'Performance benchmarks' and 'Common issues' sections into the existing references (or a new one) to make the SKILL.md body a leaner overview and improve progressive_disclosure.

Trim exclamatory commentary inside code blocks ('70B model now fits on single 24GB GPU!') to improve conciseness.

Dimension	Reasoning	Score
Conciseness	Mostly efficient, dense code and tables with little concept re-teaching, but includes minor padding such as exclamatory asides ('70B model now fits on single 24GB GPU!', 'Trains only 0.01% of parameters!') that could be trimmed.	4 / 5
Actionability	Fully executable, copy-paste-ready code spanning LoRA, QLoRA, adapter loading/merging, multi-adapter serving, IA3, prefix tuning, and TRL/vLLM integration, covering the common cases.	5 / 5
Workflow Clarity	The quick-start sequence (load → configure → train → save) is present but lacks explicit validate→fix→retry checkpoints; because fine-tuning is a batch operation, the rubric caps workflow_clarity at 3.	3 / 5
Progressive Disclosure	Well-organized sections with clearly signaled, one-level-deep references to real files (advanced-usage.md, troubleshooting.md), but benchmarks and a common-issues section that overlaps the troubleshooting reference are inlined rather than split out.	4 / 5
	Total	16 / 20 Passed

Description

87%Weight 40%Scale 1-5

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

A strong, third-person description that clearly states both the skill's purpose and concrete trigger conditions with domain-specific natural terms. Minor gains available by enumerating the '25+ methods' claim and tightening trigger phrasings.

Dimension	Reasoning	Score
Specificity	Lists several concrete capabilities (LoRA, QLoRA, '25+ methods', multi-adapter serving, 'train <1% of parameters'), but '25+ methods' is an unenumerated claim rather than discrete actions, leaving minor gaps in coverage.	4 / 5
Completeness	Explicitly answers both what ('Parameter-efficient fine-tuning for LLMs using LoRA, QLoRA, and 25+ methods') and when ('Use when fine-tuning large models (7B-70B) with limited GPU memory... or for multi-adapter serving') with concrete trigger phrases.	5 / 5
Trigger Term Quality	Good natural keyword coverage ('fine-tuning', 'LoRA', 'QLoRA', 'limited GPU memory', 'multi-adapter serving', '7B-70B') with synonyms (LoRA/QLoRA/parameter-efficient), though a few common user phrasings are absent.	4 / 5
Distinctiveness Conflict Risk	Occupies a clear niche (parameter-efficient fine-tuning via LoRA/QLoRA) with distinct, specific triggers and minimal overlap risk with other skills.	5 / 5
	Total	18 / 20 Passed

Validation

87%

Warnings & errors only

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation — 14 / 16 Passed

Validation for skill structure

Criteria	Description	Result
metadata_version	'metadata.version' is missing	Warning
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning

	Total	14 / 16 Passed

Repository: synthetic-sciences/openscience
Path: backend/cli/skills/ml-training/peft/SKILL.md
Commit: 543b17f

Reviewed: 5 days ago

Table of Contents

Discovery Implementation Validation

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.