-
Dear authors,
Thank you for opening the fantastic project!
BTW, I am wondering where I can find the negative loss of CDN.
I could see that you used the only positive cases for losses at https://g…
-
6 | DN-DAB-Deformable-DETR-R50-v24 | R50 | 49.5 (48.4 in 24 epochs) | Google Drive / BaiDu | Optimized implementation with deformable attention in both encoder and decoder. See DAB-DETR for mo…
-
Thanks for your great work!
As mentioned in Appendix D.1,
> Considering that the number of objects in one image in COCO dataset ranges from 1 to 80, this design is inefficient and results in excessi…
fushh updated
2 years ago
-
Hi,
Great work! I'm confused about why the attention mask is only used for self-attention? 如果去掉self-attention模块,只保留cross-attention模块,会不会造成noised boxes之间的信息泄露呢?
[https://github.com/IDEA-opensource/D…
-
https://github.com/IDEA-opensource/DN-DETR/blob/f41c276fe0af61a8acfbd32dfdde5d00291b3cf9/models/dn_dab_deformable_detr/dn_components.py#L103-L108
I think L105 should be
```
diff[:, :2] = known_bb…
-
I modified `projects/configs/denoise/petrv2_vovnet_gridmask_p4_800x320_dn.py` to generate a config file using ResNet-50DCN as backbone. After training, I got a lower performance compared to the paper:…
-
https://github.com/megvii-research/MOTRv2/blob/be49b7336218e470c9ebcd34be54fe7eec702675/models/motr.py#L672
当gtboxes为none或为噪声的GT, frame_res['pred_logits']值差别很大。
-
Thanks for your excellent work. I tried to apply the improved batchformerV2 code based on deformable-detr to DETR with the following modifications [here](https://github.com/OBVIOUSDAWN/tmpchange/blob/…
-
https://github.com/IDEA-opensource/DN-DETR/blob/206fa267ba7df978fa968edda9f7dd351a4b72c1/models/DN_DAB_DETR/DABDETR.py#L484
Hi thanks for wonderful repo.
Is there any reason the default num_cls = …
-
https://github.com/IDEA-opensource/DN-DETR/blob/206fa267ba7df978fa968edda9f7dd351a4b72c1/engine.py#L84-L88
It looks like you forgot optimizer.zero_grad().