Thanks for the repository.
In the appendix of your paper, it says "For the COCO-standard, we train 180k iterations, which includes 1/2/6/12/20k iterations for 0:5%/1%/2%/5%/10% in the Burn-In stage and the remaining iterations in the Teacher-Student Mutual Learning stage. ".
I notice that in this repository, all the burn in steps are set as 2k in the config files. Would that affect the results?
wjn922
commented
2 years ago
Thanks for the repository. In the appendix of your paper, it says "For the COCO-standard, we train 180k iterations, which includes 1/2/6/12/20k iterations for 0:5%/1%/2%/5%/10% in the Burn-In stage and the remaining iterations in the Teacher-Student Mutual Learning stage. ". I notice that in this repository, all the burn in steps are set as 2k in the config files. Would that affect the results?