Convert new Labelbox JSON format to YOLO

[JitskedeVries]

And add explanation about how to run

🛠️ PR Summary

_{Made with ❤️ by Ultralytics Actions}

📊 Key Changes

• Added support for converting the new Labelbox JSON format to the YOLO format.
• Implemented downloading Labelbox images directly using requests and handling network issues with retries.
• Altered the JSON parsing to accommodate the structure of Labelbox data, including extraction and normalization of bounding boxes.
• Included creation of .names file for storing unique object names sorted alphabetically.
• Added statistics on the smallest object width, height, and area encountered during conversion.

🎯 Purpose & Impact

This update enables users to convert the latest JSON format from Labelbox to YOLO format, which is crucial for training YOLO-based object detection models. By automating the download of images and the generation of corresponding annotations, it streamlines the workflow for practitioners looking to use Labelbox datasets. Furthermore, statistics about the smallest objects can help users understand dataset characteristics, potentially influencing model tuning and performance.

🌟 Summary

This PR empowers users to convert seamlessly from Labelbox's updated JSON format to YOLO format for object detection training, enhancing compatibility and dataset utilization within the Ultralytics ecosystem. 🔄💻🎯

[glenn-jocher]

@JitskedeVries to convert the new Labelbox JSON format to YOLO, you can use the 'labelbox2yolo.py' script included in the Ultralytics YOLO repos. First, ensure you have the required dependencies installed, then run the script, specifying the input JSON file and the output directory. After conversion, you can use the YOLO model for detection or training.

For running the YOLO model, please refer to the official documentation at https://docs.ultralytics.com for detailed instructions.

Let me know if you have any further questions or need additional assistance.