Dutch lemmatisation model

[proycon]

Fine-tune a dutch lemmatisation model, prime it with extra information from e-lex beforehand.

[proycon]

Possibly relevant Papers:

• Cross-Lingual Lemmatization and Morphology Tagging with Two-Stage Multilingual BERT fine-tuning 2019
• Simple Data-Driven Context-Sensitive Lemmatization 2006
• Context Sensitive Lemmatization Using Two Successive Bidirectional Gated Recurrent Networks

Possibly relevant tools:

• https://github.com/tsolakghukasyan/d-lemma
• https://docs.rs/diffs/0.4.0/diffs/myers/index.html

[proycon]

Implemented a helper tool that computes the classes for the lemmatiser: https://github.com/proycon/sesdiff