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
Hi authors, the paper mentions that an ‘end-to-end speed benchmark was established’ for a fair speed comparison, is the implementation of this benchmark included in the code in rt-detr of paddle framework?
In other words, if I train RT-DETR with improved backbone and hybrid encoders on other datasets, and it achieves 32FPS with eval, can I still directly describe it as a real-time target detector?
Sincerely, I would appreciate to have your answer!
The text was updated successfully, but these errors were encountered:
Hi authors, the paper mentions that an ‘end-to-end speed benchmark was established’ for a fair speed comparison, is the implementation of this benchmark included in the code in rt-detr of paddle framework?
In other words, if I train RT-DETR with improved backbone and hybrid encoders on other datasets, and it achieves 32FPS with eval, can I still directly describe it as a real-time target detector?
Sincerely, I would appreciate to have your answer!
The text was updated successfully, but these errors were encountered: