This repository provides a structured approach to fine-tuning the YOLOv9 model for custom object detection tasks. The included notebook and modular project structure aim to simplify the process for both beginners and advanced users.
- Step-by-Step Notebook: A complete guide to fine-tuning YOLOv9 using pre-trained weights and a custom dataset.
- Dataset Preparation: Tools to combine and process training and validation datasets dynamically.
- Training and Validation: Scripts and configurations to train and validate your YOLOv9 model effectively.
- Inference: Run inference on test images to evaluate the model’s performance visually.
- Customizable Configurations: Example
data.yamland model architecture files to get started quickly.
Below is an example of the model's inference results on the test dataset:
This project makes use of the following resources:
-
YOLOv9 Repository
- Repository: WongKinYiu/yolov9
- License: GPL-3.0 License
-
Fine-Tuning Notebook
- Author: SkalskiP
- Notebook adapted for this project. Attribution as per their repository.
-
Roboflow Dataset
- Dataset: El Señor de la Noche
- License: CC BY 4.0 License
- Changes made: Dataset combined and reformatted for YOLOv9 fine-tuning.
To get started with fine-tuning YOLOv9, download the notebook provided in this repository. The notebook contains all the necessary steps to:
- Clone the YOLOv9 repository.
- Prepare the dataset.
- Fine-tune the YOLOv9 model.
- Perform validation and inference.
-
Download the Notebook
Click here to downloadfine_tuning_yolov9.ipynb. -
Upload the Notebook to Google Colab
- Open Google Colab.
- Click on File > Upload Notebook and select the downloaded notebook.
-
Run the Notebook
Follow the step-by-step instructions in the notebook to:- Clone the YOLOv9 repository.
- Prepare and configure your dataset.
- Fine-tune the YOLOv9 model.
- Validate the model and perform inference on test images.