Train a vision transformer model to distinguish between the different grades of diabetic retinopathy (DR). DR is a disease occurring due to diabetes and can cause blindness if undetected earlier. We can limit the damage of diabetic retinopathy by early detecting and management. But in several countries (under development, especially), face difficulties in diagnosing DR due to lack of tools and the right physicians to perform the diagnosis. With the help of machine learning we can facilitate and lower the burden on physicians by performing a first diagnosis to help triage and accelerate the diagnostic process.