Fine-tuned models for extractive question answering in the Slovenian language

Fine-tuned models for extractive question answering in the Slovenian language

CLARIN.SI data & tools

Authors: Borovič, Mladen ; Žagar, Kristjan ; Ferme, Marko ; Majninger, Sandi and Ojsteršek, Milan

Item identifier: http://hdl.handle.net/11356/1778

Project URL: https://rsdo.slovenscina.eu/en/semantic-resources-and-technologies

Demo URL: https://slovenscina.eu/odgovarjanje-na-vprasanja

Date issued: 2022-09-22

Type: toolService

Language(s): Slovenian

Description: 6 different fine-tuned Transformer-based models that solve the downstream task of extractive question answering in the Slovenian language. The fine-tuned models included are: bert-base-cased-squad2-SLO, bert-base-multilingual-cased-squad2-SLO, electra-base-squad2-SLO, roberta-base-squad2-SLO, sloberta-squad2-SLO and xlm-roberta-base-squad2-SLO. The models were trained and evaluated using the Slovene translation of the SQuAD2.0 dataset (https://www.clarin.si/repository/xmlui/handle/11356/1756). The models achieve these metric values: sloberta-squad2-SLO: EM=67.1, F1=73.56 xlm-roberta-base-squad2-SLO: EM=62.52, F1=69.51 bert-base-multilingual-cased-squad2-SLO: EM=61.37, F1=68.1 roberta-base-squad2-SLO: EM=58.23, F1=64.62 bert-base-cased-squad2-SLO: EM=55.12, F1=60.52 electra-base-squad2-SLO: EM=53.69, F1=60.85

Publisher: Faculty of Electrical Engineering and Computer Science, University of Maribor

Subject(s): question answering slovenian language SQuAD extractive question answering language model

Collection(s): CLARIN.SI data & tools

Show full item record

Files in this item

This item is

Publicly Available

and licensed under:
Creative Commons - Attribution 4.0 International (CC BY 4.0)

Name: bert-base-cased-squad2-SLO.zip
Size: 388.99 MB
Format: application/zip
Description: bert-base-cased-squad2-SLO model weights and tokenizer settings
MD5: 71899094d6de396594ae2b58aa0caf94

Download file Preview

File Preview

bert-base-cased-squad2-SLO
- README.md1 kB
- tokenizer_config.json553 B
- all_results.json698 B
- pytorch_model.bin410 MB
- train_results.json197 B
- config.json722 B
- training_args.bin3 kB
- eval_nbest_predictions.json48 MB
- tokenizer.json653 kB
- vocab.txt208 kB
- special_tokens_map.json125 B
- eval_null_odds.json591 kB
- trainer_state.json7 kB
- eval_results.json503 B
- eval_predictions.json506 kB

Name: sloberta-squad2-SLO.zip
Size: 396.31 MB
Format: application/zip
Description: sloberta-squad2-SLO model weights and tokenizer settings
MD5: 094683353dc635a22f189ff7995b5750

Download file Preview

File Preview

sloberta-squad2-SLO
- README.md1 kB
- tokenizer_config.json588 B
- all_results.json692 B
- pytorch_model.bin419 MB
- train_results.json191 B
- config.json742 B
- training_args.bin3 kB
- eval_nbest_predictions.json52 MB
- tokenizer.json2 MB
- special_tokens_map.json353 B
- eval_null_odds.json591 kB
- trainer_state.json586 B
- eval_results.json503 B
- eval_predictions.json529 kB

Name: roberta-base-squad2-SLO.zip
Size: 445.73 MB
Format: application/zip
Description: roberta-base-squad2-SLO model weights and tokenizer settings
MD5: c3988fd4ba2d3c85e3f2a94a16d12c14

Download file Preview

File Preview

roberta-base-squad2-SLO
- README.md1 kB
- tokenizer_config.json1 kB
- all_results.json697 B
- pytorch_model.bin473 MB
- train_results.json197 B
- merges.txt445 kB
- vocab.json779 kB
- config.json782 B
- training_args.bin3 kB
- eval_nbest_predictions.json48 MB
- tokenizer.json2 MB
- special_tokens_map.json957 B
- eval_null_odds.json591 kB
- trainer_state.json6 kB
- eval_results.json502 B
- eval_predictions.json512 kB

Name: electra-base-squad2-SLO.zip
Size: 393.07 MB
Format: application/zip
Description: electra-base-squad2-SLO model weights and tokenizer settings
MD5: 283c80c273f7bfb2f5540315eaf387c2

Download file Preview

File Preview

electra-base-squad2-SLO
- README.md1 kB
- tokenizer_config.json534 B
- all_results.json696 B
- pytorch_model.bin415 MB
- train_results.json196 B
- config.json846 B
- training_args.bin3 kB
- eval_nbest_predictions.json49 MB
- tokenizer.json694 kB
- vocab.txt226 kB
- special_tokens_map.json125 B
- eval_null_odds.json593 kB
- trainer_state.json6 kB
- eval_results.json502 B
- eval_predictions.json549 kB

Name: xlm-roberta-base-squad2-SLO.zip
Size: 790.68 MB
Format: application/zip
Description: xlm-roberta-base-squad2-SLO model weights and tokenizer settings
MD5: 5768a0d2aee33029dabf05b81a54db4a

Download file Preview

File Preview

xlm-roberta-base-squad2-SLO
- README.md1 kB
- tokenizer_config.json451 B
- all_results.json696 B
- pytorch_model.bin1 GB
- train_results.json196 B
- config.json721 B
- training_args.bin3 kB
- eval_nbest_predictions.json50 MB
- tokenizer.json16 MB
- special_tokens_map.json280 B
- eval_null_odds.json592 kB
- trainer_state.json8 kB
- eval_results.json502 B
- eval_predictions.json538 kB

Name: bert-base-multilingual-cased-squad2-SLO.zip
Size: 636.27 MB
Format: application/zip
Description: bert-base-multilingual-cased-squad2-SLO model weights and tokenizer settings
MD5: 0d9eea42f6cda732c6a87195fb7a4782

Download file Preview

File Preview

bert-base-multilingual-cased-squad2-SLO
- README.md1 kB
- tokenizer_config.json360 B
- all_results.json695 B
- pytorch_model.bin676 MB
- train_results.json195 B
- config.json846 B
- training_args.bin3 kB
- eval_nbest_predictions.json49 MB
- tokenizer.json2 MB
- vocab.txt972 kB
- special_tokens_map.json125 B
- eval_null_odds.json591 kB
- trainer_state.json4 kB
- eval_results.json502 B
- eval_predictions.json526 kB