DridiM/INDABAX2024-Tunisia-winning-solution

# INDABAX DeepLearning 2025 Tunisia Winning solution The indabaX Tunisia is the most prestigious and challenging deep learning hackathon in Tunisia. The 2024 edition was held at Sup'com on May 18-19, and the challenge focused on object detection in dental imagery. I'm thrilled to share my solution, which achieved the best score in the competition. ## Model The model used is YOLOv8 with stratified k-folds (k = 5). ## Solution ## How to use To run tests, run the following command ```bash connect to Weights and biases with your account try: api_key = "Enter your WandB API here" wandb.login(key=api_key) except: print('If you want to use your W&B account, go to Add-ons -> Secrets and provide your W&B access token. ' 'Use the Label name as wandb_api. \nGet your W&B access token from here: wandb.ai') ``` ```bash Import libraries ``` ```bash Run the notebook. ``` ## Acknowledgements - Special thanks to indabaX and Udini for hosting this challenging hackathon .