DOFA-CLIP

Scores multispectral or RGB imagery against text descriptions, without task-specific training.

Field
Remote sensing
Method
Foundation model
Data it takes
imagery
License
CC-BY-NC-4.0 — restricted use

DOFA-CLIP pairs DOFA’s wavelength-conditioned image encoder with a text encoder, scoring imagery against a written prompt. Unlike RemoteCLIP it is not limited to RGB.

In a comparison against RemoteCLIP on a 50-chip multispectral set, the additional bands did not on their own outperform the better architecture.

The weights are non-commercial, the only such license in this catalog.

Upstream project →

Catalog entry last checked 2026-08-17. All models →

Run DOFA-CLIP 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