Platform

Volumes

Persistent storage that survives instance restarts — upload data, attach volumes to any GPU instance.

Overview

Volumes are persistent storage directories that survive instance restarts and can be attached to any GPU instance. Use them to store datasets, model checkpoints, and training outputs.

Key features:

Persistent: Data survives instance stop/start/delete
Attachable: Mount to any instance at creation time
File management: Upload, download, list, and delete files via API
Configurable size: 1-100 GB per volume

Pricing

from gpuniq import GPUniq
client = GPUniq(api_key="gpuniq_your_key")

pricing = client.volumes.pricing()
print(f"${pricing['price_per_gb_day']}/GB/day")

Create a Volume

vol = client.volumes.create(
    name="my-dataset",
    size_limit_gb=50,
    description="Training data for BERT fine-tuning",
)
print(f"Volume ID: {vol['id']}")

Parameters

body

name

Volume name. Alphanumeric, hyphens, underscores only (1-64 chars).

body

size_limit_gb

Size limit in GB (1-100).

body

description

Optional description (max 256 chars).

List Volumes

volumes = client.volumes.list()
for vol in volumes:
    print(f"{vol['name']} — {vol['used_size_bytes']} bytes used, status: {vol['status']}")

# Archived volumes
archived = client.volumes.list_archived()

Upload Files

Upload local files to a volume:

client.volumes.upload(
    volume_id=1,
    file_path="/local/path/data.tar.gz",
    subpath="datasets/",
)

Maximum upload size: 500 MB per file. For larger files, use SCP/rsync directly to the instance.

List Files

files = client.volumes.list_files(volume_id=1, subpath="datasets/")
for f in files["files"]:
    print(f"{f['name']} — {f['size']} bytes")

Download Files

# Get raw bytes
content = client.volumes.download(volume_id=1, path="datasets/data.tar.gz")

# Save directly to local file
client.volumes.download_to(
    volume_id=1,
    remote_path="datasets/data.tar.gz",
    local_path="./data.tar.gz",
)

Delete Files

client.volumes.delete_file(volume_id=1, path="datasets/old_data.tar.gz")

Update Volume

client.volumes.update(volume_id=1, size_limit_gb=100, description="Updated description")

Delete Volume

client.volumes.delete(volume_id=1)

Deleting a volume permanently removes all files. This action is irreversible.

Attach Volume to Instance

Pass volume_id when creating an order:

# Marketplace order with volume
order = client.marketplace.create_order(
    agent_id=123,
    pricing_type="hour",
    volume_id=9,
)

# Dex-Cloud deploy with volume
deploy = client.gpu_cloud.deploy(
    gpu_name="RTX_4090",
    volume_id=9,
)

# Burst order with volume
order = client.burst.create_order(
    docker_image="pytorch/pytorch:latest",
    primary_gpu="RTX_4090",
    gpu_count=4,
    volume_id=9,
)

Sync Logs

Monitor volume sync operations:

logs = client.volumes.sync_logs(volume_id=1, limit=50)
client.volumes.cancel_sync(log_id=5)

Was this page helpful?

Last updated Feb 22, 2026

Built with Documentation.AI