분석시각화 대회 코드 공유 게시물은
내용 확인 후
좋아요(투표) 가능합니다.
합성데이터 기반 객체 탐지 AI 경진대회
Is YOLOv5 allowed?
Is it allowed to use the YOLOv5 (https://github.com/ultralytics/yolov5) and other GPL/AGPL licensed models in this competition? Thanks in advance.
yolov5는 논문이 존재하지 않는걸로 보이는데 사전학습모델 사용이 불가능하나요?
Hello @DACONIO,
Thank you for your response. Could I ask clarifying questions?
1) What counts as a paper?
According to this discussion (https://github.com/ultralytics/yolov5/issues/1333) YOLOv5 seemingly does not have a dedicated paper (e.g. on Arxiv). But there is a Zenodo publication (https://doi.org/10.5281/zenodo.3908559). Also there are several papers by different authors which describe the model:
[1] https://www.theseus.fi/bitstream/handle/10024/452552/Do_Thuan.pdf,
[2] https://www.researchgate.net/publication/357684232_Comparing_YOLOv3_YOLOv4_and_YOLOv5_for_Autonomous_Landing_Spot_Detection_in_Faulty_UAVs,
[3] https://www.bib.irb.hr/1217017/download/1217017.Article.pdf,
[4] https://arxiv.org/abs/2108.11539
Do some of these publications allow us to use the model in the competition?
2) Is the AGPL license compatible with the winner's obligations?
AGPL license allows all kinds of usage (https://github.com/ultralytics/yolov5/blob/master/LICENSE), but the whole code of the solution must be sub-licensed under AGPL too. Is it suitable?
Thanks in advance!
Hello vecxoz,
I would like to respond to your message.
1) YOLOv5 is considered as a model based on model published in a paper,
so it is allowed to use in this competition
2) AGPL license does compatible with winner’s obligation
Thank you
안녕하세요, 이용성 님
YOLOv5는 사용 가능하십니다.
감사합니다.
Hello @DACONIO,
This is great! Thank you very much.
我对你的指责有同感。
肖,你什么意思?
SSD object detection 사전학습된 모델도 사용가능한가요?? 빠른 답변 부탁 드립니다 ㅠㅠ
사용 가능합니다.
데이콘(주) | 대표 김국진 | 699-81-01021
통신판매업 신고번호: 제 2021-서울영등포-1704호
서울특별시 영등포구 은행로 3 익스콘벤처타워 901호
이메일 dacon@dacon.io | 전화번호: 070-4102-0545
Copyright ⓒ DACON Inc. All rights reserved
Hello vecxoz,
If there are no legal restrictions on using pre-trained model that you are intending to use,
and if the model is based on a model published in a paper,
then it would be available for your use.
For further information, please check "4. External data and pre-trained model" under following link.
https://dacon.io/en/competitions/official/236107/overview/rules
Thank you