The FaQuAD dataset is a reading comprehension dataset designed for evaluating question-answering models. Let me provide you with details about different versions of FaQuAD:
-
FaQuAD (English):
- Description: The original FaQuAD dataset follows the format of the Stanford Question Answering Dataset (SQuAD).
- Content: It comprises 900 questions related to 249 reading passages. These passages were extracted from 18 official documents of a computer science college at a Brazilian federal university and 21 Wikipedia articles related to the Brazilian higher education system³.
- Purpose: Researchers use FaQuAD to develop and evaluate reading comprehension models in the domain of Brazilian higher education.
- GitHub Repository: You can find the dataset and related code on the FaQuAD GitHub repository.
-
FQuAD (French):
- Description: FQuAD is a French Native Reading Comprehension dataset created by higher education students. It consists of 25,000+ questions based on a set of Wikipedia articles.
- Similarity to SQuAD: Like SQuAD, FQuAD provides annotated questions and answers for evaluation purposes.
- Website: You can explore the FQuAD dataset on the FQuAD website.
-
FaQuAD (Portuguese):
- Description: As far as we know, FaQuAD is a pioneer Portuguese reading comprehension dataset that follows the challenging format of SQuAD.
- Source: The dataset includes passages from official documents of a Brazilian computer science college and Wikipedia articles related to Brazilian higher education.
- GitHub Repository: The FaQuAD data and source code for experiments are available on the FaQuAD GitHub repository.
In summary, FaQuAD provides valuable resources for training and evaluating question-answering models across different languages and domains. Researchers can use these datasets to advance natural language understanding and improve machine comprehension systems.
Source: Conversation with Bing, 3/16/2024
(1) FaQuAD: Reading Comprehension Dataset in the Domain of ... - ResearchGate. https://www.researchgate.net/profile/Eraldo-Fernandes/publication/337789791_FaQuAD_Reading_Comprehension_Dataset_in_the_Domain_of_Brazilian_Higher_Education/links/5e825f5fa6fdcc139c173c8f/FaQuAD-Reading-Comprehension-Dataset-in-the-Domain-of-Brazilian-Higher-Education.pdf.
(2) GitHub - liafacom/faquad: FaQuAD reading comprehension dataset and .... https://github.com/liafacom/faquad.
(3) FQuAD. https://fquad.illuin.tech/.
(4) ruanchaves/faquad-nli · Datasets at Hugging Face. https://huggingface.co/datasets/ruanchaves/faquad-nli.
(5) FaQuAD: Reading Comprehension Dataset in the Domain of ... - GitHub. https://github.com/liafacom/faquad?search=1.