eduagarcia/sick-br

---
annotations_creators:
- expert-generated
language_creators:
- found
language:
- pt
license:
- unknown
multilinguality:
- monolingual
size_categories:
- 10K<n<100K
source_datasets:
- original
task_categories:
- text-classification
- sentence-similarity
task_ids:
- text-scoring
- natural-language-inference
- semantic-similarity-scoring
pretty_name: SICK-BR
dataset_info:
features:
- name: pair_ID
dtype: string
- name: sentence_A
dtype: string
- name: sentence_B
dtype: string
- name: entailment_label
dtype: string
- name: relatedness_score
dtype: float64
- name: entailment_AB
dtype: string
- name: entailment_BA
dtype: string
- name: sentence_A_original
dtype: string
- name: sentence_B_original
dtype: string
- name: sentence_A_dataset
dtype: string
- name: sentence_B_dataset
dtype: string
splits:
- name: train
num_bytes: 1341470
num_examples: 4439
- name: validation
num_bytes: 151133
num_examples: 495
- name: test
num_bytes: 1480190
num_examples: 4906
download_size: 840022
dataset_size: 2972793
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*
- split: test
path: data/test-*
---

# SICK-BR

This is a reproduction of the of the [SICK-BR dataset](https://github.com/livyreal/SICK-BR) in HuggingFace format, SICK-BR is a Portuguese inference corpus, aligned to and translated from SICK (Marelli et al., 2014).

To know more about the resource, please see the paper: [Real, Livy; Rodrigues, Ana; Vieira e Silva, Andressa; Albiero, Beatriz; Thalenberg, Bruna; Guide, Bruno; Silva, Cindy; Oliveira Lima, Guilherme de, Camara, Igor C. S.; Stanojevic, Milos; Souza, Rodrigo; Paiva, Valeria de. SICK-BR: a Portuguese corpus for inference. PROPOR2018 (International Conference on the Computational Processing of Portuguese), 26 September 2018. Canela, Brazil](https://linux.ime.usp.br/~thalen/SICK_PT.pdf)

Part of the SICK-BR was used in ASSIN 2, a shared task of inference and semantic similarity held in Salvador (BA/Brazil), in October 2019, in conjuction with STIL 2019.
https://sites.google.com/view/assin2/

## Citation

```bibtex
@inproceedings{real18,
author="Real, Livy
and Rodrigues, Ana
and Vieira e Silva, Andressa
and Albiero, Beatriz
and Thalenberg, Bruna
and Guide, Bruno
and Silva, Cindy
and de Oliveira Lima, Guilherme
and C{\^a}mara, Igor C. S.
and Stanojevi{\'{c}}, Milo{\v{s}}
and Souza, Rodrigo
and de Paiva, Valeria"
year ="2018",
title="SICK-BR: A Portuguese Corpus for Inference",
booktitle="Computational Processing of the Portuguese Language. PROPOR 2018.",
doi ="10.1007/978-3-319-99722-3_31",
isbn="978-3-319-99722-3"
}
```

## Dataset Description

- **Homepage:** https://github.com/livyreal/SICK-BR
- **Paper:** [SICK-BR: A Portuguese Corpus for Inference](https://linux.ime.usp.br/~thalen/SICK_PT.pdf)

原始信息汇总

SICK-BR 数据集概述

基本信息

• 语言: 葡萄牙语 (pt)
• 许可证: 未知
• 多语言性: 单语
• 数据集大小: 10K<n<100K
• 数据来源: 原始数据

任务类型

• 文本分类
• 句子相似度

具体任务

• 文本评分
• 自然语言推理
• 语义相似度评分

数据集特征

• pair_ID: 字符串
• sentence_A: 字符串
• sentence_B: 字符串
• entailment_label: 字符串
• relatedness_score: 浮点数
• entailment_AB: 字符串
• entailment_BA: 字符串
• sentence_A_original: 字符串
• sentence_B_original: 字符串
• sentence_A_dataset: 字符串
• sentence_B_dataset: 字符串

数据分割

• 训练集: 4439 个样本，大小为 1341470 字节
• 验证集: 495 个样本，大小为 151133 字节
• 测试集: 4906 个样本，大小为 1480190 字节

下载与数据集大小

• 下载大小: 840022 字节
• 数据集总大小: 2972793 字节