maelicn

3 ms·

1. 1.▲
   by maelicn 2mo ago

   The first (and only) model for predicting any relations between any regions (boxes, masks) of an image. It runs localy on your browser or on a GPU under 20ms. This model is my biggest release so far, made entirely on my free time during the
2. 2.▲
   Show HN: RelateAnything – Real-time open-vocabulary relation prediction(github.com)

   2 points by maelicn 2mo ago|1 comments