Open dkyman opened 10 months ago
Hi @dkyman,
It wouldn't hurt to check with them, however the author(s) of the upstream for this project released their code under the MIT license (see LICENSE file). This is a permissive license which allows us to modify and redistribute with minimal restriction, as long as our version contains the same license. That also means that all code under this project is open source under the same permissive terms.
We're still fixing some issues with this repo for patch training with YOLOv8 (#2), please check back with us soon though!
We probably should clarify the terms of the components we added to this project repo.
The Airbus dataset is under CC-SA 4, which is generally compatible with MIT.
The yolov8 model we trained (best.pt
) using transfer learning from the pretrained COCO weights is most likely AGPL-3.0, since that weights file we trained from was in a repo marked as AGPL-3.0. Only the file best.pt
would be under the restriction imposed by this permissive copyleft license.
Hi @dkyman,
It wouldn't hurt to check with them, however the author(s) of the upstream for this project released their code under the MIT license (see LICENSE file). This is a permissive license which allows us to modify and redistribute with minimal restriction, as long as our version contains the same license. That also means that all code under this project is open source under the same permissive terms.
We're still fixing some issues with this repo for patch training with YOLOv8 (#2), please check back with us soon though!
Dear author, can I understand that you will soon open source the engineering code for YOLOv8 patches?
This is the code, it's currently a work in progress
Could I ask the author if you can open source the code for the YOLOv8 adverse patch