ModernGBERT DROC Mention Tagger

This model is a sequence tagger and a fine-tuned version of LSX-UniWue/ModernGBERT_134M on the DROC coreference dataset to annotate mentions of characters in literary text. Training and testing was performed on 1024 token windows. It achieves the following results on the evaluation document Die Wahlverwandtschaften, annotated in the context of the GerFuN corpus:

  • Loss: 0.0334
  • Precision: 0.9577
  • Recall: 0.9665
  • F1: 0.9620
  • Accuracy: 0.9907
Downloads last month
13
Safetensors
Model size
0.1B params
Tensor type
BF16
·
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for aehrm/moderngbert-droc-tagger

Finetuned
(9)
this model