adversarial-attack A basic gradio based app to study how the models predictions vary across various adversarial attacks. To run: uvicorn main:app --reload python3 gradio-ui.py TODO: Add more attacks Add other models Improve visual inference