Sonata
Creates embeddings from 3D point clouds without requiring labelled training data.
- Field
- Cross-domain
- Method
- Foundation model
- Data it takes
- point-cloud
- License
- Apache-2.0 code, CC-BY-NC-4.0 weights — restricted use
Sonata is a self-supervised encoder: it learns point-cloud representations from unlabelled scans, and those embeddings then feed a classifier trained on a small labelled set.
This is the point-cloud counterpart to what Clay does for satellite imagery, and it suits the same situation — a segmentation or classification task where labelling at scale is not affordable.
The code is Apache-2.0 but the weights are non-commercial, and the weights are what binds a user.
Catalog entry last checked 2026-08-19. All models →
Run Sonata on your data
The lab runs this on its own compute, in a container, with the inputs and parameters recorded alongside the result. Initial scoping conversations are free.
Contact the Lab