Skip to content

Model benchmarking / leaderboard population #47

Description

@LasseQV

8/10
currently have scripts for qwen and gemma for detection
need to switch to yaml based config to allow for comments
need to check qwen image scaling, see if it is possible to re-enable dynamic scaling and get bounding boxes that seem reasonable
next step is to add a few other open source vlms
also add vlm scripts for classification

then make scripts for open vocabulary detection (grounding dino, yolo-world, etc)

Metadata

Metadata

Assignees

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