Fine-tuning

For conceptual background on fine-tuning methods, training requirements, and when to use each approach, see Fine-tuning.

Fine-tuning workflows

Create a fine-tuning job

Set up projects, configure training and infrastructure, and launch a fine-tuning job.

Reinforcement tuning

Train reasoning models with reward-based optimization using GRPO.

Preference tuning

Align model outputs with human preferences using DPO.

Vision language tuning

Fine-tune vision-language models on image-text datasets.

Low-rank adaptation

Apply low-rank adaptation to reduce training cost and memory usage.

Deploy a fine-tuned model

Promote models for inference, run chat completions, and manage deployments.