File size: 2,816 Bytes
473d856 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 |
---
license: apache-2.0
language:
- sr
metrics:
- accuracy
base_model:
- classla/bcms-bertic
library_name: transformers
---
# BERTić-COMtext-SR-legal-lemma-ijekavica
**BERTić-COMtext-SR-legal-lemma-ijekavica** is a variant of the [BERTić](https://huggingface.co/classla/bcms-bertic) model, fine-tuned on the task of lemmatization tag prediction in Serbian legal texts written in the Ijekavian pronunciation.
The model was fine-tuned for 20 epochs on the Ijekavian variant of the [COMtext.SR.legal](https://github.com/ICEF-NLP/COMtext.SR) dataset.
# Benchmarking
This model was evaluated on the task of lemmatizing Serbian legal texts.
Lemmatization was performed using the predicted string edit tags, as described in this JTDH 2024 paper:
* [Lemmatizing Serbian and Croatian via String Edit Prediction](https://zenodo.org/records/13937204)
The model was compared to previous lemmatization approaches that relied on the [hrLex](http://hdl.handle.net/11356/1232) inflectional lexicon:
- The [CLASSLA](http://pypi.org/project/classla/) library
- A variant of [BERTić](https://huggingface.co/classla/bcms-bertic) fine-tuned for MSD prediction using the [SETimes.SR 2.0](http://hdl.handle.net/11356/1843) corpus of newswire texts
- A [variant](https://huggingface.co/ICEF-NLP/bcms-bertic-comtext-sr-legal-msd-ijekavica) of [BERTić](https://huggingface.co/classla/bcms-bertic) fine-tuned for MSD prediction using the [COMtext.SR.legal](https://github.com/ICEF-NLP/COMtext.SR) corpus of legal texts
- [SrBERTa](http://huggingface.co/nemanjaPetrovic/SrBERTa), a model specially trained on Serbian legal texts, fine-tuned for MSD prediction using the [COMtext.SR.legal](https://github.com/ICEF-NLP/COMtext.SR) corpus of legal texts
Accuracy was used as the evaluation metric and gold tokenized text was taken as input.
All of the previous large language models were fine-tuned for 15 epochs.
CLASSLA and BERTić-SETimes were directly tested on the entire COMtext.SR.legal.ijekavica corpus.
BERTić-COMtext-SR-legal-MSD-ijekavica, BERTić-COMtext-SR-legal-lemma-ijekavica, and SrBERTa were fine-tuned and evaluated on the COMtext.SR.legal.ijekavica corpus using 10-fold CV.
The code and data to run these experiments is available on the [COMtext.SR GitHub repository](https://github.com/ICEF-NLP/COMtext.SR).
## Results
| Model | Lemma ACC |
| ------------------------------------------- | ---------- |
| CLASSLA-SR | 0.9036 |
| CLASSLA-HR | 0.9353 |
| BERTić-SETimes.SR | 0.9412 |
| BERTić-COMtext-SR-legal-MSD-ijekavica | 0.9429 |
| SrBERTa | 0.9187 |
| **BERTić-COMtext-SR-legal-lemma-ijekavica** | **0.9833** |
|