Beam allows you to create highly-available storage volumes that can be used across tasks. You might use volumes for things like storing model weights or large datasets.

​
Mounting a Volume

Volumes can be attached to an App. By default, Volumes are shared across all apps in your Beam account.

from beam import App, Volume

app = App(
    name="my_app",
    volumes=[
        Volume(
            name="my_models",
            path="models",
        )
    ],
)

You can also restrict a volume to a specific app by specifying a VolumeType when your volume is registered to an app: volume_type=VolumeType.Persistent


from beam import App, Volume, VolumeType

app = App(
    name="my_app",
    volumes=[Volume(name="my_models", path="models", VolumeType=VolumeType.Persistent)],
)

​
Reading and writing to volumes

You can read and write to your Volume like any ordinary Python file:

from beam import App, Volume

volume_path = "./models"

app = App(
    name="my_app",
    volumes=[
        Volume(
            name="my_models",
            path=volume_path,
        )
    ],
)


@app.run()
def write_files():
    with open(f"{volume_path}/somefile.txt", "w") as f:
        f.write("Writing to the volume!")


@app.run()
def read_files():
    with open(f"{volume_path}/somefile.txt", "r") as f:
        f.read()

​
Uploading Files with the CLI

You can upload files with the Beam CLI, for example:

beam volume upload [volume name] [file path] -a [app name]

If uploading to a Persistent Volume type, the app name must be provided using the -a flag:

beam volume upload my_volume some_file.png -a the_app

If uploading to a Shared Volume, the app flag can be excluded.

​
Uploading Files with the Dashboard

You can also upload files with the web dashboard, on the volumes page.

​
Uploading Folders Recursively in Python

You can recursively upload folders through the CLI, or through a Python script:

1

First Step

Copy the code below to a file on your local machine. Save the file as upload.py.

2

Second Step

Run this command:

beam run upload.py:upload -d '{"local_path": "./YOUR_DIR", "beam_volume_path": "./YOUR_BEAM_VOLUME"}'

3

Third Step

Run beam volume ls [YOUR_VOLUME_NAME] to verify that your files were uploaded!

​
Browsing Volume Contents

You can list all your volumes through the web dashboard, or CLI using the beam volume list command.

You can also browse all contents of a volume with the beam volume ls command:

​
Deleting Volumes

You can delete a volume in the web dashboard. Click the three-dots icon next to a volume, and click Delete Volume to permanently delete it.

Was this page helpful?

Configuring Keep WarmTask Outputs