Getting YOLOv8 detection output as sigmoid

Krisequ commented 7 months ago

Search before asking

[X] I have searched the YOLOv5 issues and discussions and found no similar questions.

Question

Hey,

I am working on a custom training approach for detectional neural networks (using YOLOv8). Please correct me if I am wrong, but typically, after finding anchors, the task changes to classification and linear regression of the bounding box size. These typically are done with sigmoid functions. Using basic Ultralytics functions like predict/train, the framework provides BBOX in pixels, which is perfectly understandable. I was trying to access the values as sigmoid values (0 to 1), but I seem to be unable to find it at any stage of the processing.

Where can I find a function that translates sigmoids to BBOX coordinates (so I could reverse it at a later stage), or is there a way to access the sigmoid outputs from the network, at best also during the training?

Best regards

Additional

No response

github-actions[bot] commented 7 months ago

👋 Hello @Krisequ, thank you for your interest in YOLOv5 🚀! Please visit our ⭐️ Tutorials to get started, where you can find quickstart guides for simple tasks like Custom Data Training all the way to advanced concepts like Hyperparameter Evolution.

If this is a 🐛 Bug Report, please provide a minimum reproducible example to help us debug it.

If this is a custom training ❓ Question, please provide as much information as possible, including dataset image examples and training logs, and verify you are following our Tips for Best Training Results.

Requirements

Python>=3.8.0 with all requirements.txt installed including PyTorch>=1.8. To get started:

git clone https://github.com/ultralytics/yolov5  # clone
cd yolov5
pip install -r requirements.txt  # install

Environments

YOLOv5 may be run in any of the following up-to-date verified environments (with all dependencies including CUDA/CUDNN, Python and PyTorch preinstalled):

Notebooks with free GPU:
Google Cloud Deep Learning VM. See GCP Quickstart Guide
Amazon Deep Learning AMI. See AWS Quickstart Guide
Docker Image. See Docker Quickstart Guide

Status

If this badge is green, all YOLOv5 GitHub Actions Continuous Integration (CI) tests are currently passing. CI tests verify correct operation of YOLOv5 training, validation, inference, export and benchmarks on macOS, Windows, and Ubuntu every 24 hours and on every commit.

Introducing YOLOv8 🚀

We're excited to announce the launch of our latest state-of-the-art (SOTA) object detection model for 2023 - YOLOv8 🚀!

Designed to be fast, accurate, and easy to use, YOLOv8 is an ideal choice for a wide range of object detection, image segmentation and image classification tasks. With YOLOv8, you'll be able to quickly and accurately detect objects in real-time, streamline your workflows, and achieve new levels of accuracy in your projects.

Check out our YOLOv8 Docs for details and get started with:

pip install ultralytics

Krisequ commented 7 months ago

Only now I found this thread: https://github.com/ultralytics/ultralytics/issues/6982 and realized that there's different approach for detection in yolov8 than I expected.

glenn-jocher commented 7 months ago

@Krisequ hello!

You're correct that YOLOv8, like its predecessors, uses a different approach for bounding box prediction. The network outputs are typically transformed from raw model outputs to bounding box coordinates through a series of operations, including the application of the sigmoid function to certain outputs to constrain them between 0 and 1.

If you're looking to work with the raw sigmoid outputs, you would need to modify the inference code to intercept these values before they are converted to bounding box coordinates. This would involve diving into the model's forward pass and extracting the relevant tensors.

For reversing the process (i.e., converting bounding box coordinates back to the network's raw outputs), you would need to understand the exact transformations applied and invert them. This can be quite complex due to the nature of the transformations, which include scaling and translating based on anchor boxes.

For more detailed guidance on the model architecture and output transformations, please refer to our documentation at https://docs.ultralytics.com/yolov5/.

github-actions[bot] commented 6 months ago

👋 Hello there! We wanted to give you a friendly reminder that this issue has not had any recent activity and may be closed soon, but don't worry - you can always reopen it if needed. If you still have any questions or concerns, please feel free to let us know how we can help.

For additional resources and information, please see the links below:

Docs: https://docs.ultralytics.com
HUB: https://hub.ultralytics.com
Community: https://community.ultralytics.com

Feel free to inform us of any other issues you discover or feature requests that come to mind in the future. Pull Requests (PRs) are also always welcomed!

Thank you for your contributions to YOLO 🚀 and Vision AI ⭐

ultralytics / yolov5