MLOps: Industrialize your computer vision models using Docker and REST API

Tutorial goal: Deploy a Docker image featuring a REST API built with FastAPI in order to exploit a computer vision model responsible for classifying input photographs into one of seven forms of skin cancer. The first step here is to train your own model on a specific task. In our example, we have trained an image classifier using multiple architectures (ViT, VGG16, ResNet50, DenseNet121, …) taken from TorchVision and HuggingFace to identify the type of skin cancer from the input images. If you haven't trained your own model yet, and you don't know how, you should start. I sincerely recommend you to take a look on our tutorial using HugsVision.