DataTune Client

DataTune is a unified platform for AI data workflows that simplifies how you manage, process, and stream training data for machine learning models.

The client sdk allows you to connect with the platform via code and continue with your data workflows with ease.

Website • Documentation • Blog • Quick Start• Contact Us

Key Features

• Unified Data Management: Centralize your datasets from multiple sources in one place
• Efficient Streaming: Stream data in batches directly to your machines
• Simple Integration: Get started with just a few lines of code
• Intuitive Dataset Views: Easy-to-use interface for data labeling and editing.
• Parallel processing Support: Scale your data processing with parallel workers

Platform

Head over to https://datatune.ai/ to request access to the platform.

Installation

pip install datatune-client

Quick Start

from datatune.api import API
from datatune.entity import Entity
from datatune.workspace import Workspace
from datatune.streaming import DataTuneLoader

# Initialize
api = API(api_key="your-api-key")
entity = Entity(id="your-org-id", api=api)
workspace = Workspace(entity=entity, name="your-workspace")

# Load view and start streaming
view = workspace.load_view("your-view-name")
dataloader = DataTuneLoader(view, batch_size=32, num_workers=4)

# Stream data
for batch in dataloader.stream():
    # Process your batch
    pass

Documentation

For detailed documentation and guides, visit docs.datatune.ai

Contact Us

• Email: hello@vitalops.ai
• Twitter: https://x.com/datatune_ai

License

This project is licensed under the MIT License - see the LICENSE file for details.