Skip to content

Add dedicated 'audio' key for multimodal models #650

@HelloCheck0840

Description

@HelloCheck0840

The Problem:
Currently, to pass audio data to multimodal models like gemma4:e2b, people are forced to use the images key.

Why this is an issue:

  • Naming Confusion: Currently, audio data often has to be passed into a field called images. This is confusing because audio files are not images.
  • Future Multimodal Support: For models that support both image and audio simultaneously, a single images bucket creates ambiguity.

Suggested Solution:
Add dedicated 'audio' key.

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions