Visualize encoder-decoder multi-head attention weights

IDEA-Research / DINO

[ICLR 2023] Official implementation of the paper "DINO: DETR with Improved DeNoising Anchor Boxes for End-to-End Object Detection"

Apache License 2.0

2.15k stars 232 forks source link

Visualize encoder-decoder multi-head attention weights #130

Open owen24819 opened 1 year ago

owen24819 commented 1 year ago

Are you able to visualize encoder-decoder multi-head attention weights with the deformable attention similar to what was done in the original DETR paper? I know the Deformable DETR paper was able to do this but they did not post their code or respond to issues on their repo.