Using YOLO correctly?

RoboCupJunior Rescue RoboCupJunior Rescue Line
1

Hello,

I have a question about if it’s permissible to use YOLO, since I’ve read many posts in which some people say yes and others say no. In fact, I saw this same doubt in a previous post where the answer wasn’t very clear.

We are using YOLOv11n. We manually labeled our dataset with LabelImg, then split the images into training and validation sets. Finally, we compressed the dataset and uploaded it to Google Colab, where we trained the model using code based on this YouTube tutorial (original notebook: “https://colab.research.google.com/github/EdjeElectronics/Train-and-Deploy-YOLO-Models/blob/main/Train_YOLO_Models.ipynb#scrollTo=8bbpob1gTPlo”). That link describes a procedure almost identical to ours (from step 1 to 2.1, and from 3 to 5.2).

Given that we created our own dataset from scratch, we assumed our usage was allowed—but after reading conflicting opinions, we’re not so sure. Is our method of using YOLO permitted?

It’s our first time using YOLO, so we are not experts, and a detailed explanation of what is and isn’t permitted would be helpful.

Thank you in advance, and greetings,

Team Disprosio.

1 Like
2

I made a mistake. Actually i use YOLOv8n (instead of YOLOv11n) and in the step 5.2 (Run Training!) i replace “model=yolov8n.pt” by “model=yolov8n.yaml”, because I discovered that a “.pt” file contains pretrained weights, and a “.yaml” file defines only the model’s architecture and initializes random weights for a truly “from‑scratch” training.

Now, I’d like to confirm: is this allowed under the competition rules?

Thank you again,

Team Disprosio.