IDEA-Research / Stable-DINO

[ICCV 2023] Official implementation of the paper "Detection Transformer with Stable Matching"
https://arxiv.org/abs/2304.04742
Apache License 2.0
218 stars 7 forks source link

About alpha in Position-Supervised Loss #7

Open Iranb opened 1 year ago

Iranb commented 1 year ago

In the paper, you change the classification target from [0, 1] to [0, $f_1(s_i)$]. Does the alpha setting during training change or not in your focal loss?