search

nicholasmaplin / cvat-segmentation-mask-to-yolo

this code will create the text files that are needed to train YOLOv8 seg models when images labelled using SAM inside CVAT
1 stars 0 forks source link

readme

YOLOv8seg Annotation File Converter

This repository contains a Python script that converts annotation files generated by CVAT (using the SAM model) into the required format for training a YOLOv8seg model. The script creates .txt files containing bounding box coordinates and class labels.

Usage

  1. Clone the Repository:

    git clone https://github.com/nicholas.m.aplin/cvat-segmention-mask-to-yolo.git
cd cvat-segmention-mask-to-yolo

  2. Install Dependencies: Install the required packages:

    pip install -r requirements.txt

  3. Run the Script: Execute the convert_annotations.py script, passing the path to your CVAT annotation file as an argument:

    python cvat_to_yolo.py --input json_file_name.json, /path/to/folder

  4. Output: The script will generate .txt files in the YOLO format (one per image) containing bounding box coordinates and class labels.

Example

Check out the example.py file in this repository for a usage example.