Train AI model to detect images

We want to develop a model, and train it to detect images on a Dermatology database. Nvidia documentation says we do not have to start from scratch. Can you recommend an existing model we can use to save time in developing our dermatology model? any sample work would help. I noticed under similar topics, "creating medical imaging models with Nvidia Clara Train 4.0, never used it, so not sure. We want to have a patient take a skin image on their smart phone and identify that image from the use of the model.