This is me and my friends' project submission to Hack The North 2021. This repo is the machine learning model part, namely the initial Mark 1 product.
We trained a object recognition model specialized in detecting cigarette butts using a dataset from Kaggel and YOLOv5. This model can be used in environment clean up robots as a key component.
The trained results is in mark_1_results.zip
To re-create the results download the dataset from here and the two jupyter notebooks to your google colab and run them.
reference: