Labelformat - Fast Label Conversion for Computer Vision

Labelformat is an open-source Python framework for converting between popular computer vision annotation formats like YOLO, COCO, PascalVOC, and KITTI. Save hours on tedious format conversions and ensure consistency in your workflows.

Key Features

• Wide Format Support: COCO, YOLO (v5-v11), PascalVOC, KITTI, Labelbox, and more.
• Cross-Platform: Compatible with Python 3.7+ on Windows, macOS, and Linux.
• Flexible Usage: Intuitive CLI and Python API.
• Efficient: Memory-conscious, optimized for large datasets.
• Offline First: Operates locally without data uploads.
• Tested for Accuracy: Round-trip tests for consistent results.

Get Started Quickly

1. Install via pip:

   pip install labelformat
   

2. Convert Labels in One Command:

   labelformat convert --task object-detection \
                       --input-format coco \
                       --input-file coco-labels/train.json \
                       --output-format yolov8 \
                       --output-file yolo-labels/data.yaml
   

Supported Formats

2D Object Detection Label Formats

Format	Read ✔️	Write ✔️
COCO	✔️	✔️
KITTI	✔️	✔️
Labelbox	✔️	❌
Lightly	✔️	✔️
PascalVOC	✔️	✔️
YOLOv5 - v11	✔️	✔️

---

2D Instance Segmentation Label Formats

Format	Read ✔️	Write ✔️
COCO	✔️	✔️
YOLOv8	✔️	✔️

Explore More

• Quick Start Guide
• Detailed Usage Guide
• List of all features

---

📦 Quick Links

• GitHub Repository
• PyPI Package
• Documentation

Labelformat is maintained by Lightly.