Skip to content

Add support for text-only and multi-image vlm eval #144

Description

@camilobrownpinilla

What

Extend the lmms-eval harness to support text-only and multi-image vlm benchmarking .

Scope

This should only alter the adapter.

Backward compatibility

This will be fully backwards compatible.

Metadata

Metadata

Labels

enhancementNew feature or request

Type

No type

Fields

No fields configured for issues without a type.

Projects

No projects

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions