> ## Documentation Index
> Fetch the complete documentation index at: https://docs.harborframework.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Datasets

> Collections of tasks for evaluation and training.

A [Harbor task](/tasks/overview.md) is an instruction, sandbox environment, and test script. A dataset is a collection of tasks. Datasets sometimes define custom metrics that aggregate task rewards.

There are three ways to use datasets:

1. **Local datasets**: run a local directory of tasks or `dataset.toml` manifest
2. **Published datasets**: run a dataset from [Harbor Hub](https://hub.harborframework.com/)
3. **Git repository datasets**: run a dataset from a Git repository

## Local datasets

Run a local dataset with `--path` or `-p`:

```bash theme={"system"}
harbor run -p "<path/to/dataset>" -a "<agent>" -m "<model>"
```

The path can be a directory of tasks or a `dataset.toml` manifest file.

## Published datasets

Datasets can be published and shared with members of your organization or publicly on [Harbor Hub](https://hub.harborframework.com/).

Run a published dataset with `--dataset` or `-d`:

```bash theme={"system"}
harbor run -d "<org/name@version>" -a "<agent>" -m "<model>"
```

To learn how to create and publish a dataset, see [Publish a dataset](/datasets/create-a-dataset.md#publish-a-dataset).

## Git repository datasets

Run a dataset directly from any Git repository with `--repo`:

```bash theme={"system"}
harbor run --repo org/repo-name -p "./tasks" -a "<agent>" -m "<model>"
```

Supports GitHub, GitLab, and Hugging Face URLs, with optional ref pinning (`@v1.0`, `@main`).

See [Git repos](/datasets/git-repos.md) for the full guide.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.