Issues with training the Yolov8 (all) OBB model when importing the dataset from Roboflow
Search before asking
- [X] I have searched the HUB issues and found no similar bug report.
HUB Component
Training
Bug
Issues with training the Yolov8 (all) OBB model when importing the dataset from Roboflow Training proceeds normally, but at the validation stage, it does not go further, neither in the cloud version nor in the Colab version when training through Hub. There are also path issues in the local version, and it doesn't proceed past validation. On local training has this issue: FileNotFoundError: Dataset 'D://ML/hub/ObjectDetectionOBB.v4i.yolov8-obb/data.yaml' images not found โ ๏ธ, missing path 'D:\ML\datasets\roboflow\valid\images' Note dataset download directory is 'D:\ML\hub'. You can update this in 'C:\Users\vladk\AppData\Roaming\Ultralytics\settings.yaml'
When training the same dataset for the normal version, without OBB, everything works fine.
A few questions: How can You fix cloud training? How can I adjust the settings for proper operation in local training?
Environment
Ultralytics HUB Version v0.1.51 Client User Agent Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/128.0.0.0 Safari/537.36 Operating System Win32 Browser Window Size 2560 x 1305 Server Timestamp 1726828671
Minimal Reproducible Example
- Download my demo obb dataset with 20 pictures from Roboflow as obb v 8 zip https://universe.roboflow.com/pregmate/objectdetectionobb/dataset/5
- Import as OBB v8
- Try train by any way
Additional
No response
๐ Hello @VladKovalski, thank you for raising an issue with the Ultralytics HUB repository ๐! An Ultralytics engineer will assist you soon. In the meantime, let's make sure we have all the details we need to address your issue.
It seems you're experiencing problems during the validation stage when training the YOLOv8 OBB model, both on the cloud and locally. For the local training issue, there's a FileNotFoundError indicating a path problem. Please ensure the dataset paths are correctly set in your settings.yaml as mentioned.
To assist us further, could you please provide:
- A detailed description or screenshot of the error you're encountering.
- Confirmation that the dataset paths are correctly configured in your settings.yaml.
For both cloud and local training:
- Minimum Reproducible Example (MRE): Please include a clear, minimal example that reproduces the issue. It seems you've started on thisโthanks for the steps provided!
If you're looking for more detailed guidance on configuring your setup or exporting models, you might find our HUB Docs helpful:
- Datasets: Preparing and Uploading. Learn about proper dataset preparation, which could aid in resolving path issues.
- Models: Training and Exporting. This section covers training nuances that might be useful for your problem.
We appreciate your patience as we work through this! ๐
@VladKovalski Hello!
It seems the dataset format is incorrect, and you'll need to adjust it. The Roboflow-exported dataset may have been exported in a different format.
Unfortunately, we've encountered issues with the Roboflow format. For example: https://github.com/ultralytics/hub/issues/835.
@sergiuwaxmann Hi. Thank you very much for your response, but I have a problem specifically with the OBB data. I labeled them according to the Roboflow service documentation and exported them in Yolo v8 OBB format. I can record a video of the process. How can I properly label the data for OBB or correctly upload it from Roboflow?
๐ 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 ๐ and Vision AI โญ