A visual transformer model based on google/vit-base-patch16-224-in21k to detect if a gym progress pic was taken from the front, from the back, from the left side or from the right side. There is a demo, but really this is just my first dummy project using the transformers library.