CREDENCE
Collection
CREDENCE credal concept-bottleneck models for CEBaB uncertainty experiments. • 51 items • Updated
Paper: https://huggingface.co/papers/2604.24170 · arXiv: 2604.24170
This checkpoint is packaged with PyTorchModelHubMixin.
Install the dependencies and clone Tankiit/Credal_Sets, then run:
from models.slvm_base import HybridCredalSLVM
model = HybridCredalSLVM.from_pretrained("thomasbllx/credence-cebab-aleprior-p0.05")
The original legacy checkpoint and evaluation metrics are included for reproducibility.