Skip to content

Making deep learning backends truly optional at import time #168

Description

@C-Achard

Right now, importing DLCLive also imports torch/tf in several places without guards.

Examples

Examples include :

Whereas certain modules such as utils.py properly guard the imports.

Motivation

Properly guarded imports and ideally introducing "lazier" behavior would drastically help in making truly lightweight tests in CI, and likely improve import times drastically.

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

Labels

dependenciesPull requests that update a dependency fileenhancementNew feature or requestpytorchRelated to the pytorch backendtensorflow

Type

Projects

No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions