implementing-llms-litgpt

Implements and trains LLMs using Lightning AI's LitGPT with 20+ pretrained architectures (Llama, Gemma, Phi, Qwen, Mistral). Use when need clean model implementations, educational understanding of architectures, or production fine-tuning with LoRA/QLoRA. Single-file implementations, no abstraction layers.

Quality

78%

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

Fix and improve this skill with Tessl

tessl review fix ./backend/cli/skills/ml-training/litgpt/SKILL.md

Quality

Content

68%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 highly actionable skill body with strong file structure and reference navigation, weakened by redundant LoRA examples, verbose restated checklists, and missing validation checkpoints in the batch training workflows.

Suggestions

Add explicit validation checkpoints to the training workflows (e.g., after fine-tuning, evaluate on a held-out set and confirm the checkpoint loads before merging or deploying).

De-duplicate the LoRA content: Workflow 1's finetune_lora example and Workflow 2's overlap heavily — keep one canonical example and reference it, or move the detailed parameter guide to a reference file.

Trim the checkbox checklists that restate the immediately-following step headers, or collapse them so the steps are not stated twice.

Dimension	Reasoning	Score
Conciseness	Mostly useful specifics without over-explaining basic concepts, but carries real redundancy: near-identical finetune_lora blocks appear in both Workflow 1 and Workflow 2, and the checkbox checklists restate the step headers that immediately follow.	3 / 5
Actionability	Fully executable, copy-paste-ready commands and code cover the common cases (download, finetune, finetune_lora, merge_lora, pretrain, quantize, GGUF conversion, FastAPI deploy) with real flags and parameters.	5 / 5
Workflow Clarity	The four workflows are clearly sequenced with checklists, but training and fine-tuning are batch operations with no explicit validation checkpoints (e.g., eval/verify model quality before proceeding), so the batch-operation cap at 3 applies.	3 / 5
Progressive Disclosure	Advanced topics points to four well-signaled one-level-deep references (all files exist) and offloads exhaustive lists, but the body is heavier than an overview with inline detail (full pretrain config, duplicate LoRA examples) that could be split out.	4 / 5
	Total	15 / 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, specific description that clearly states what the skill does and when to use it, anchored to a named library and concrete techniques. Minor weaknesses are slightly broad action verbs and one abstract trigger phrase.

Dimension	Reasoning	Score
Specificity	Lists several concrete actions ("Implements and trains LLMs", "production fine-tuning with LoRA/QLoRA") plus named architectures, but core actions like "trains LLMs" are broader than the anchor-5 level of concrete operations.	4 / 5
Completeness	Explicitly answers both what ("Implements and trains LLMs using Lightning AI's LitGPT with 20+ pretrained architectures") and when ("Use when need clean model implementations... production fine-tuning with LoRA/QLoRA") with concrete trigger phrases.	5 / 5
Trigger Term Quality	Good natural keyword coverage ("fine-tuning", "LoRA/QLoRA", model family names users cite), but misses some common synonyms and the phrase "educational understanding of architectures" is a goal rather than something a user says.	4 / 5
Distinctiveness Conflict Risk	Names a specific library (LitGPT) with a clear niche and distinctive triggers (LoRA/QLoRA, single-file implementations), giving minimal conflict 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
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning
referenced_paths_exist	Referenced path issues: 2 missing	Warning

	Total	14 / 16 Passed

Repository: synthetic-sciences/openscience
Path: backend/cli/skills/ml-training/litgpt/SKILL.md
Commit: 3a6c3a9

Reviewed: 1 day 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.