Fine-Tuning Details about your proposal #665
-
|
Good morning. I have been reading your proposal and would like to cite it in one of my papers. However, I would like to clarify some details. I have seen that it is contemplated the re-training applying SAHI. How does it work exactly, do you re-train with the outputs obtained after applying YOLOv5+SAHI or do you use the dataset dedicated to Visdrone training already annotated previously? On the other hand, I would like to get the json files to generate the mAPs set in the experimental phase according to your ablation study. |
Beta Was this translation helpful? Give feedback.
Replies: 1 comment
-
|
Hello @IvanGarcia7, thanks for your interest in SAHI! Let's continue the paper related discussion here: https://github.com/fcakyon/sahi-benchmark/issues/1 |
Beta Was this translation helpful? Give feedback.
Hello @IvanGarcia7, thanks for your interest in SAHI!
Let's continue the paper related discussion here: https://github.com/fcakyon/sahi-benchmark/issues/1