Serverless Fine-Tuning

Train open models for your domain or task—without provisioning or managing GPU infrastructure yourself—using serverless fine-tuning. After you evaluate the training results, use self-serve deployments to deploy your fine-tuned model to a dedicated inference endpoint.

How it worksConcepts, lifecycle, and architecture behind serverless fine-tuning.QuickstartRun your first fine-tuning job end-to-end in a few minutes.Fine-tune a modelConfigure datasets, hyperparameters, and checkpoints for a production run.Self-serve deploymentsDeploy a fine-tuned checkpoint to a dedicated inference endpoint.