Art style classifier and gallery

The brief

An online art gallery wants visitors to discover paintings they will like. Build a model that recognises art style and artist, and use image embeddings to recommend similar paintings. Good looks like: good style accuracy, recommendations a designer agrees with, and a browsable gallery.

What you will build

1. Explore styles, genres and artists
2. Fine-tune an image model for style
3. Build image embeddings for similarity
4. Create a recommendation gallery
5. Reflect on what the model gets wrong about art

Free, open data to build it with

• WikiArt ↗ Hugging Face

Check each dataset's licence before you publish your work. Your mentor gives your team its own version of the brief (data, scope or region), so every team's project is its own.