Model weights for a study of commonsense reasoning

Handle:	https://hdl.handle.net/21.11129/0000-000F-4869-B (persistent URL to this page)
URL:	https://github.com/nlx-group/study-of-commonsense-reasoning

This resource contains model weights for five Transformer-based models: RoBERTa, GPT-2, T5, BART and COMET(BART).
These models were implemented using HuggingFace, and fine-tuned on the following four commonsense reasoning tasks: Argument Reasoning Comprehension Task (ARCT), AI2 Reasoning Challenge (ARC), Physical Interaction Question Answering (PIQA) and CommonsenseQA (CSQA).

Download

DistributionLicence

MIT

Contact Person

Ruben Branco

University of Lisbon, Faculty of Sciences

FCUL

[javascript protected email address]

Department of Informatics

http://nlx.di.fc.ul.pt/

FCUL

Faculdade de Ciências de Lisboa, Departamento de Informática. Campo Grande, 1749-016 Lisboa, Portugal

1749-016 Lisbon

Portugal

[javascript protected email address]

Tel.: +351 217 500 087

Fax: +351 217 500 084

Language Description
text

Language Description General Information

Other

Monolingual text languageDescriptionLanguages

English

Linguality

Linguality type: Monolingual

Metadata

Created: 09/09/2021

Last Updated: 11/23/2021

Metadata Creator

João Ricardo Silva

http://nlx-server.di.fc.ul.pt/~jsilva/

University of Lisbon, Faculty of Sciences

FCUL

[javascript protected email address]

Sala 6.3.32, Edifício C6, Departamento de Informática, Faculdade de Ciências da Universidade de Lisboa

1749-016 Lisboa

Campo Grande

Portugal

Department of Informatics

http://nlx.di.fc.ul.pt/

FCUL

Faculdade de Ciências de Lisboa, Departamento de Informática. Campo Grande, 1749-016 Lisboa, Portugal

1749-016 Lisbon

Portugal

[javascript protected email address]

Tel.: +351 217 500 087

Fax: +351 217 500 084

People who looked at this resource also viewed the following: