Integrations

Kangas integrates with a few popular ML libraries and platforms. If there is a library you'd like Kangas to integrate with that you can't find below, please feel free to open a ticket. We prioritize integrations in order of community demand, so your feedback is very helpful.

Hugging Face

Deploy

The Kangas Python library can deploy hosted instances of Kangas to Hugging Face Spaces, prepopulated with your DataGrid. This is achieved with the kangas.integrations.huggingface.deploy_to_huggingface method:

def deploy_to_huggingface(path, name):
    """
    Deploy DataGrid as part of a hosted Kangas instance on
    HuggingFace Spaces. Overwrites any existing DataGrids with
    identical names.

    Args:
        path: (str) Name of repository to create
        name: (str) DataGrid to use
    """

Example:

from kangas.integrations.huggingface import deploy_to_huggingface

deploy_to_huggingface(path="YOUR-HF-SPACE", name="YOUR-DATA.datagrid")

Import/Export

The Kangas CLI can import/export Hugging Face Datasets directly. For more details, see https://github.com/comet-ml/kangas/wiki/Kangas-CLI#kangas-importexport

Comet

Import/Export

The Kangas CLI can import/export Comet Experiments directly. For more details, see https://github.com/comet-ml/kangas/wiki/Kangas-CLI#kangas-importexport