This Is the implementation of the first of the two parts of the Huawei challenge @HackaTUM22 In this repo you will find the our solution to the image binary classification task. Our final solution is using a Visual Transformer reaching 99.2 validation accuracy. To easily test the model go at: https://huggingface.co/ogimgio/vit-base-highways-2 where you will be able to upload pictures and get the prediction on the go. In the notebooks you can also see our CNN attempt.