YAML Metadata Warning:empty or missing yaml metadata in repo card
Check out the documentation for more information.
Classes
- Asphalt
- Tiles
- Gravel
- Unpaved
- Green roof
- Greenhouse
- Concrete surface
- Cobblestones
- Solar panels
Primary use case
The primary use case of these models is to distinguish between paved and unpaved surfaces.
The paved category is further subdivided into material-specific classes, such as asphalt, tiles, concrete, cobblestones, etc.
Production setup
In production, the models are run as a 3-model ensemble.
Performance:
- Pixels correctly classified: 97.94%
Lightweight single-model option
If a single, lightweight model is preferred, use the ResNet34-based model.
Performance on our validation set:
- Pixels correctly classified: 97.1%
- Pixels correctly classified as paved vs. unpaved: 99.7%
- Pixels with label
unpavedcorrectly classified: 99.83% - Pixels with paved labels incorrectly classified as
unpaved: 1.9%
license: mit
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support