Mechanism

When It Triggers · How It Works · What It Produces

Breaks down how the skill works and what it produces, so you can quickly judge whether it fits your scenario and value.

Trigger

  • Triggered when users need to fine-tune models (SFT/DPO/RFT).
  • Covers data preparation, training submission, grader calibration, model deployment, and evaluation sub-scenarios.

Mechanism

  • Golden rules: always baseline first → validate data → calibrate RFT graders → evaluate checkpoints → measure token cost vs accuracy.
  • 11 Python scripts covering the full pipeline: data conversion, validation, training submission, monitoring, grader calibration, deployment, evaluation, distillation data generation, cleanup.
  • Multiple workflow guides and reference files covering quickstart, full pipeline, dataset creation, iterative training, and diagnostics.

Outcome

  • Fine-tuned model deployed to Foundry with evaluation report (accuracy, token cost vs baseline). Checkpoints can be selected for replacement.

Use Cases

Two Typical Use Cases

Scenario 01

End-to-End Model Fine-Tuning Automation

AI engineers need to fine-tune base models for domain-specific tasks. The skill supports SFT/DPO/RFT training types, starting from data preparation: formatting training data, running validators for quality checks, submitting training jobs, monitoring progress, analyzing training curves, then deploying and evaluating fine-tuned models. Complete script toolchain provided.

Scenario 02

RFT Grader Calibration and Iterative Training

When using RFT (reinforcement with graders), the grader's pass_threshold must be precisely calibrated to achieve 25-50% failure rate on the base model. The skill provides calibration scripts, supports iterative training, analyzes training curves and evaluates checkpoints to select the best model, avoiding blind deployment of the final checkpoint.

Skill Relationships

Which Skills I Use · Which Skills Use Me

Dependency relationships read as "the upper tier points to the lower tier." The current Skill sits in the middle tier — above are Skills that depend on it, below are Skills it depends on.

Tier 1 · These Skills Use Me

Tier 2 · Current Skill

  • finetuningWorkflow

Tier 3 · I Use These Skills

Full Relationship Graph

Browse this skill's relationships within its skillset. Click a node to switch the side panel; use the search box to jump to any skill.

Skill File

SKILL.md

Open in GitHub ↗
SKILL.mdLoading from GitHub…