FeverAlbert

FeverAlbert is a classifier model that predicts if evidence is related to query claim. The model achieved F1 score of 88.33% with test dataset "mwong/fever-claim-related". Using pretrained albert-base-v2 model, the classifier head is trained on Fever dataset.

Downloads last month
11
Inference Providers NEW
This model is not currently available via any of the supported Inference Providers.

Dataset used to train mwong/albert-base-fever-claim-related