This project is in beginning phase.

Welcome to Semantix, a GUI-based open-source tool designed to simplify the process of labeling and training data for object detection using the YOLO (You Only Look Once) algorithm. Semantix streamlines the workflow of preparing a dataset, training the model, and generating predictions.

• Easy-to-use, user-friendly GUI.
• Supports .jpg and .png image files.
• Allows users to load images and corresponding labels with a single click.
• Facilitates dataset preparation in YOLO format.
• Provides options to set key training parameters like epochs and learning rate.
• Integrated support for multiple CUDA devices.
• Displays real-time training logs in the console.
• Automatic saving of weight files in the YOLO file system upon completion of training.
• Enables label generation for a large set of images.
• Supports Semantic Segmentation output with polygons (or rectangles) for different object classes.

Clone this repository and navigate into it:

git clone https://github.com/yourgithubusername/Semantix.git
cd Semantix

Install the required dependencies:

pip install -r requirements.txt

Run the main script:

python main.py

This project is licensed under the terms of the MIT License. For more details, see the LICENSE file.