full-form of "c" and "e" in yolov9 models

[0xSynapse]

We can understand that this yolov9 models are sorted according to the param-count and performance.

But what do you mean by c and e in yolov9-c and yolov9-e model names?

[ashish-roopan]

They have not mentioned that in the paper. I would be nice to know that.

[WongKinYiu]

c and e indicate compact and extended models, which are mentioned in Table 3 and Section 5.2, respectively.

[JohnMBrandt]

Adding onto this, it doesn't seem that the # params matches the paper. Torch tells me that yolov9-c has 51043480 parameters and yolov9-e has 69437144 parameters -- Nevermind, I see this is due to the aux branch

[WongKinYiu]

Aux branch has additional parameters and can be removed at inference stage. We provide re-parameterization script to get the converted model for inference. You could download yolov9-c-converted.pt to see real #params and FLOPs.

[0xSynapse]

Thank you @WongKinYiu, I have published a blog explaining YOLOv9 for OpenCV at LearnOpenCV's official site. You can give it a read YOLOv9: Advancing the YOLO Legacy --- If you agree, we can add this to the YOLOv9 readme file.

[WongKinYiu]

Added to readme.

[0xSynapse]

@WongKinYiu I just added a PR with an updated button for the blog, as it's difficult for viewers to find.

I hope you will see it.

Thank you.