Need help with rotate-yolov3?
Click the “chat” button below for chat support from the developer who created it, or find similar developers for support.

About the developer

181 Stars 49 Forks 27 Commits 26 Opened issues


Arbitrary oriented object detection implemented with yolov3 (attached with some tricks).

Services available


Need anything else?

Contributors list

# 22,245
27 commits


Rotaion object detection implemented with yolov3.

Hello, the no-program ryolov3 is available now. Although not so many tricks are attached like this repo, it still achieves good results, and is friendly for beginners to learn, have a good luck.


The latest code has been uploaded, unfortunately, due to my negligence, I incorrectly modified some parts of the code and did not save the historical version last year, which made it hard to reproduce the previous high performance. It is tentatively that there are some problems in the loss calculation part.

But I found from the experimental results left last year that yolov3 is suitable for rotation detection. After using several tricks (attention, ORN, Mish, and etc.), it have achieved good performance. More previous experiment results can be found here.


  • SEBlock
  • riou loss
  • Inception module
  • DCNv2
  • ORN
  • SeparableConv
  • Mish/Swish
  • GlobalAttention

Detection Results

The detection results from rotated yolov3 left over last year:


Following questions are frequently mentioned. And if you have something unclear, don't doubt and contact me via opening issues.

  • Q: How can I obtain


sets the initial anchors generated via kmeans, you need to run
refer to my implemention here. You can also check
for more details.
  • Q: How to train the model on my own dataset?

A: This ryolo implemention is based on this repo, training and evaluation pipeline are the same as that one do.

  • Q: Where is ORN codes?

A: I'll release the whole codebase as I return school, and this repo may help.

  • Q: I cannot reproduce the result you reported(80 mAP for hrsc and 0.7 F1 for IC15).
  • A: Refer to my reply here. This is only a backup repo, the overall model is no problem, but direct running does not necessarily guarantee good results, cause it is not the latest version, and some parameters may have problems, you need to adjust some details and parameter settings yourself. I will upload the complete executable code as soon as I return to school in September (if lucky).

In the end

There is no need or time to maintain the codebase to reproduce the previous performance. If you are interested in this work, you are welcome to fix the bugs in this codebase, and the trained models are available here with extracted code

. I'll reimplement the rotation yolov4 or yolov5 if time permitting in the future.

We use cookies. If you continue to browse the site, you agree to the use of cookies. For more information on our use of cookies please see our Privacy Policy.