You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Thanks for this amazing work! I have some question about adding DN to a Vallina-DETR like model.
Could you explain more about how can I use DN for a Vallina-DETR like algorithm?
Because the Vallina-DETR's object quries are not anchor-like, thus I don't know how to change an denoised gt(dim=a) to a obj query(dim=b && b != a)?
Will a learnable nn.linear or other oporator work?
Looking forward to your reply!
The text was updated successfully, but these errors were encountered:
Thanks for this amazing work! I have some question about adding DN to a Vallina-DETR like model.
Could you explain more about how can I use DN for a Vallina-DETR like algorithm?
Because the Vallina-DETR's object quries are not anchor-like, thus I don't know how to change an denoised gt(dim=a) to a obj query(dim=b && b != a)?
Will a learnable nn.linear or other oporator work?
Looking forward to your reply!
The text was updated successfully, but these errors were encountered: