Search before asking
Description
i am using your notebook Automated Dataset Generation with Grounding DINO + Segment Anything Model (SAM)
on google colab
on some simple example from street views, Grounding Dino labels a car both as a "car" and as a "truck" though truck as a lower threshold.
i can play on one image on the box-threshold but i can't if i want to use the approach to do auto labelling on >100s of images.
how can i keep only the max threshold for same bounding boxes?
Additional
No response
Are you willing to submit a PR?