metadata
github: https://github.com/3D-fossils-Haag/nsm
datasets:
- BioVisionLab/vertebrae_meshes_v72
- BioVisionLab/train_data
base_model:
- BioVisionLab/run_v72
tags:
- NSM
- DeepSDF
- TriplanarDecoder
- SignedDistanceFunction
- generativeAI
- 3D
- lizard
- vertebrae
- squamate
- reptile
- museum
language:
- en
license: openrail
Model Description
Trained on 2896 presacral vertebrae meshes from 95 specimens representing all families of lizards and a few snakes. Trained using same data and parameters as baseline v72 model, but with hierarchy loss.
Model Use
Can be used to do shape completion and classification of unknown squamate vertebrae (modern or fossil).
Citation
If you use this model, its weights, or the underlying research in your work, please cite it using the following: Wolcott et al. 2026. “Vert-NSM” GitHub repository. https://github.com/3D-fossils-Haag/vert-nsm (accessed YYYY-MM-DD).