# Network volumes

> Persistent, shareable storage for datasets, models and checkpoints.

This page shows how to create a network volume, mount it, check how much of it is used, resize it and delete it.

A network volume is persistent storage that lives outside any instance. Its data stays when instances stop or are deleted, and several instances and Run Jobs can mount it at once. See [Storage](/storage) to compare it with the container disk.

## Create a volume

<tabs sync="interface">
<tabs-item label="Console">

On the [Storage page](https://console.cloud.featherless.ai/storage), click **Create volume**, enter a name and a capacity, and click **Create volume**.

You can also create one while you deploy an instance: under **Network volumes**, click **Add network volume** and choose **New network volume…** in the list.

</tabs-item>

<tabs-item label="CLI">

```bash
featherless-cloud volume create my-data --gb 500 --wait
```

</tabs-item>
</tabs>

A volume is billed for its full capacity whether or not you use it, so start with what you need. A new volume takes about a minute to be ready for files.

## Mount a volume

Mount volumes when you deploy an instance or submit a Run Job. They can't be added to or removed from an existing instance.

<tabs sync="interface">
<tabs-item label="Console">

On the deploy page, under **Network volumes**, click **Add network volume**, choose the volume and enter the path to mount it at. The default is `/workspace`.

</tabs-item>

<tabs-item label="CLI">

```bash
featherless-cloud instance create --name dev --template featherless/rocm-pytorch --gpu-count 1 --volume my-data:/workspace
featherless-cloud exec --gpu-count 1 --volume my-data:/workspace -- python /workspace/train.py
```

</tabs-item>
</tabs>

- **Mount at /workspace** to keep your whole working directory across stops.
- **Mount at another path, such as /data,** to keep datasets and models while scratch work stays on the container disk.

An instance or Run Job that mounts a volume runs in the volume's location. Several of them can mount the same volume; give each its own output folder so two never write the same file at once.

## Check how much space is used

The [Storage page](https://console.cloud.featherless.ai/storage) shows each volume's usage and the instances that mount it. Inside an instance:

```bash
df -h /workspace
```

From the CLI, `featherless-cloud volume get my-data` prints the capacity, the usage and the instances using it.

## Resize a volume

You can increase a volume's capacity at any time, even while it's mounted, and running instances see the new size right away. You can't decrease it. Use **Edit volume** in the volume's actions menu on the Storage page, or:

```bash
featherless-cloud volume update my-data --gb 1000
```

## Delete a volume

Deleting a volume erases its data, and it can't be undone. If instances mount it, stop them first: **Detach and delete** removes the volume from them, and they start again without it.

```bash
featherless-cloud volume delete my-data
```

## Next steps

<card-group>
<card icon="i-lucide-folder-up" title="Transfer files" to="/storage/transfer-files">

Upload and download files.

</card>

<card icon="i-lucide-notebook-pen" title="Develop in JupyterLab" to="/examples/jupyterlab">

Keep notebooks on a network volume.

</card>
</card-group>
