finetuning
Official sourceFine-tunes models on Azure AI Foundry with SFT, DPO, or RFT, covering dataset prep and validation, grader calibration, training, checkpoint selection, deployment, and evaluation.
Source: Microsoft
At a glance
Verified source- Best for
- Running SFT, DPO, or RFT fine-tuning jobs on Azure AI Foundry
- Works with
- GitHub Copilot · Claude Code
- Outcome
- Preparing, validating, and converting training datasets
- Source
- microsoft/github-copilot-for-azure
01 / Understand the skill
About finetuning
Fine-tune models on Azure AI Foundry using SFT (supervised), DPO (preference), or RFT (reinforcement with graders). Covers dataset preparation, training job submission, deployment, and evaluation. USE FOR: fine-tune, SFT, DPO, RFT, training data, grader, distillation, fine-tuned model, training job, large file upload, calibrate grader, deploy fine-tuned model, evaluate fine-tuned model. DO NOT USE FOR: general model deployment without fine-tuning (use deploy-model), agent creation (use agents), prompt optimization without training (use prompt-optimizer).
Fine-tune models on Azure AI Foundry using SFT (supervised), DPO (preference), or RFT (reinforcement with graders). Covers dataset preparation, training job submission, deployment, and evaluation.
Not a fit when
DO NOT USE FOR: general model deployment without fine-tuning (use deploy-model), agent creation (use agents), prompt optimization without training (use prompt-optimizer).
Read full description
Fine-tune models on Azure AI Foundry using SFT (supervised), DPO (preference), or RFT (reinforcement with graders). Covers dataset preparation, training job submission, deployment, and evaluation. USE FOR: fine-tune, SFT, DPO, RFT, training data, grader, distillation, fine-tuned model, training job, large file upload, calibrate grader, deploy fine-tuned model, evaluate fine-tuned model. DO NOT USE FOR: general model deployment without fine-tuning (use deploy-model), agent creation (use agents), prompt optimization without training (use prompt-optimizer).
02 / Confirm the fit
Use finetuning when you need to
- Running SFT, DPO, or RFT fine-tuning jobs on Azure AI Foundry
- Preparing, validating, and converting training datasets
- Calibrating RFT graders and pass thresholds
- Evaluating and deploying fine-tuned model checkpoints
03 / See the workflow
How finetuning works
- 01Install: npx skills add https://github.com/microsoft/github-copilot-for-azure --skill finetuning
- 02Try: Run an SFT fine-tune of gpt-4.1-mini, validate my data first, then evaluate the result.
- 03Get: A baselined, validated training run with curve analysis, checkpoint choice, and eval results.
Check before installing
Compatible platforms
GitHub Copilot · Claude Code
Best fit
Running SFT, DPO, or RFT fine-tuning jobs on Azure AI Foundry
Review the source instructions if the required output differs from the formats described above.
Install and use
Compatible with GitHub Copilot and Claude Code · Published by Microsoft
Ready to add finetuning?
Use the published instructions below, then provide the required input and describe the result you need.
General instructions · GitHub Copilot · Claude Code
npx skills add https://github.com/microsoft/github-copilot-for-azure --skill finetuning
Review the source repository and license before installing third-party skills.
Continue exploring
Similar skills
Discover and install agent skills from the open ecosystem via the npx skills CLI — when you wish you had a skill for some task.
End-to-end Azure AI Foundry agent lifecycle with azd—scaffold, deploy, invoke, evaluate, optimize, fine-tune, and troubleshoot agents—routing to specialized sub-skills per task.
Unified Azure OpenAI deployment router that detects intent and dispatches to quick preset, full customize, or capacity-discovery modes, with shared project and quota validation.
Discovers available Azure OpenAI model capacity across all accessible regions and projects, then validates subscription quota and ranks deployment locations.