DataTune Client#

SDK for Datatune, a unified platform for AI data workflows.

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

Blogs#

Simplifying LLM Training with DataTune: A Beginner’s Guide

Learn how to get started with DataTune for LLM training workflows.

https://medium.com/@abhijithneilabraham/simplifying-llm-training-with-datatune-a-beginners-guide-4492c6ca5812
Simplify Large Data Processing with DataTune

Explore techniques for handling and processing large datasets efficiently with DataTune.

https://medium.com/@abhijithneilabraham/simplify-large-data-processing-with-datatune-01ecfb736bcd

SDK Examples#

Jupyter Notebook Examples

Explore practical examples of using the DataTune SDK.

https://github.com/vitalops/datatune/blob/main/examples/sdk_example.ipynb

Documentation#

License#

MIT License