Closed Mr-Bigworth closed 1 year ago
We will consider releasing this code in the future. The implementation is quite similar to DN-DETR, where we append additional GT-denoised queries to the head and train the model to reconstruct the GT.
We will consider releasing this code in the future. The implementation is quite similar to DN-DETR, where we append additional GT-denoised queries to the head and train the model to reconstruct the GT.
Thanks!!!
Excuse me, the paper (https://arxiv.org/abs/2203.01305) shows the DN training can also be used in Faster R-CNN, I wonder how to implement it, Thanks!