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

# Models

> Browse the model catalog and deploy or fine-tune a model.

Source: https://docs.zettabyte.com/docs/models-data-storage/models

Last updated: 2026-07-21

The **Models** tab in the **Library** is the model catalog. A **Pre-Trained Model** / **Custom Model** toggle switches between the provided catalog and your own models. Pre-trained models appear as cards showing an icon, name, description, and creation date, each with **Deploy** and **Fine Tuning** actions.

## Pre-trained models

The catalog includes open-weight instruction and chat models. Open the **Models** tab for the current list, since the catalog changes over time.

## Deploy a model

Select **Deploy** on a model to open the deployment wizard with the model preselected. The deployment runs as a dedicated inference endpoint; see Deploy a model.

## Fine-tune a model

Select **Fine Tuning** on a model to open the fine-tuning wizard with the model preselected:

- **Basic Information**: a **Task Name**, a **Description**, and the **Model** (pre-filled, with version).
- **Training Configuration**: a **Fine-tuning Method** (such as LoRA) and a parameter table of training hyperparameters with adjustable defaults.
- **Resource**: a **Computing Config** and **Environment Variables**. Select a launched resource (see GPU instances).
- **Dataset Configuration**: a **Training Set** and a **Validation Set**, each chosen from your datasets (see Datasets).
- **Overview**: review and create.

A simpler fine-tuning form is also available under Training and tasks.

## Custom models

The **Custom Model** view holds models you add yourself.