YOLO Object Detection Back

In my Summer 2024 internship, I used YOLOv8 and YOLOv10 to train object detection and object segmentation models. I labeled images using MakeSenese AI and trained the YOLO models in AWS. During this project, I built different methods to adress class imbalance, utilizing under and oversampling methods, preprocessing, and splitting datasets. Addtionaly, I aided in the transition from local training to cloud-based training in AWS, built our first segmentation model using YOLO, and first random forrest model with scikit-learn. I designed an ran our model benchmark testing, and predicted and recorded how different hyperparameter configurations affected model performance.