droc-singletons

Maverick Coref Model DROC

This is the Maverick corefernce model trained with German literary texts (specifically DROC).

To use it install pip install maverick-coref-de

Now running the model is as easy as:

from maverick_de import Maverick
model = Maverick(
  hf_name_or_path="uhhlt/droc-singletons",
  device="cuda",
)

print(model.predict("Es war einmal eine kleine süße Dirne, die hatte jedermann lieb, der sie nur ansah, am allerliebsten aber ihre Großmutter, die wußte gar nicht, was sie alles dem Kinde geben sollte."))

If you make use of this model in your work we would love for you to cite our paper:

@inproceedings{petersen-frey-etal-2025-efficient,
    title = "Efficient and Effective Coreference Resolution for {G}erman",
    author = "Petersen-Frey, Fynn  and
      Hatzel, Hans Ole  and
      Biemann, Chris",
    editor = "Wartena, Christian  and
      Heid, Ulrich",
    booktitle = "Proceedings of the 21st Conference on Natural Language Processing (KONVENS 2025): Long and Short Papers",
    month = sep,
    year = "2025",
    address = "Hannover, Germany",
    publisher = "HsH Applied Academics",
    url = "https://aclanthology.org/2025.konvens-1.12/",
    pages = "128--137"
}