> See the full index of documentation pages at https://docs.zettabyte.com/llms.txt.

# Fine-tune a model

> Adapt an existing model to your data with the guided fine-tuning flow.

Source: https://docs.zettabyte.com/docs/run-on-gpus/fine-tune-a-model

Last updated: 2026-07-21

## Before you begin

- A base `Model`. See Models.
- A `Training Set` and a `Validation Set` from your Datasets. See Datasets.
- Allocated capacity, and a `PVC` for outputs. See GPU instances and Storage and PVCs.

## Create a fine-tuning task

1. Open **Tasks** and use **Create Task** then **Fine-Tuning**.
2. Name the task.
3. Set the `Computing Config`: `Resource Name`, `GPU Count`, `CPU`, and `Memory`.
4. Select the `Model`, the `Training Set`, and the `Validation Set`.
5. Attach `Storage` for your outputs.
6. **Review and Create**.

> [!NOTE]
> **Need more control?**
> Fine-tuning is a specialized form of a training job. For a custom command, execution frameworks, or multi-node runs, use a `Training` task instead. See Run a training job.

## Next steps

Deploy the fine-tuned model as an inference endpoint. See Deploy a model.