NERDA 0.0.19

Named Entity Recognition for DAnish based on Transformers

NERDA

NERDA is not only a mesmerizing muppet-like character. NERDA is also a python package, that offers a complete framework for fine-tuning pretrained huggingface transformers for Named Entity Recognition (=NER) tasks.

Installation guide

pip install NERDA

NER tasks

Named Entity Recognition (NER) tasks are all about identifying and extracting names of named entitites from natural language texts.

Read more about NER on Wikipedia.

Performance

The table below summarizes the performance (=F1-scores) of the model configurations, that NERDA ships with.

Level	MBERT	DABERT	ELECTRA	XLMROBERTA	DISTILMBERT
B-PER	0.92	0.93	0.92	0.94	0.89
I-PER	0.97	0.99	0.97	0.99	0.96
B-ORG	0.68	0.79	0.65	0.78	0.66
I-ORG	0.67	0.79	0.72	0.77	0.61
B-LOC	0.86	0.85	0.79	0.87	0.80
I-LOC	0.33	0.32	0.44	0.24	0.29
B-MISC	0.73	0.74	0.61	0.77	0.70
I-MISC	0.70	0.86	0.65	0.91	0.61
AVG_MICRO	0.81	0.85	0.79	0.86	0.78
AVG_MACRO	0.73	0.78	0.72	0.78	0.69

AVG_ stands for micro- and macro AVeraGed F1-scores.

'NERDA'?

'NERDA' originally stands for 'Named Entity Recognition for DAnish'. However, this is somewhat misleading, since the functionality is no longer limited to Danish. On the contrary it generalizes to all other languages, i.e. NERDA supports fine-tuning of transformer-based models for NER tasks for any arbitrary language.

Read more

The documentation for NERDA including code references and examples can be accessed here.

Contact

We hope, that you will find NERDA useful.

Please direct any questions and feedbacks to us!

If you want to contribute (which we encourage you to), open a PR.

If you encounter a bug or want to suggest an enhancement, please open an issue.