datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
squad
Dataset Card for SQuAD
Dataset Summary
Stanford Question Answering Dataset (SQuAD) is a reading comprehension dataset, consisting of questions posed by crowdworkers on a set of Wikipedia articles, where the answer to every question is a segment of text, or span, from the corresponding reading passage, or the question might be unanswerable.
SQuAD 1.1 contains 100,000+ question-answer pairs on 500+ articles.
Supported Tasks and Leaderboards
Question… See the full description on the dataset page: https://huggingface.co/datasets/rajpurkar/squad.squad_v2
Dataset Card for SQuAD 2.0
Dataset Summary
Stanford Question Answering Dataset (SQuAD) is a reading comprehension dataset, consisting of questions posed by crowdworkers on a set of Wikipedia articles, where the answer to every question is a segment of text, or span, from the corresponding reading passage, or the question might be unanswerable.
SQuAD 2.0 combines the 100,000 questions in SQuAD1.1 with over 50,000 unanswerable questions written adversarially by crowdworkers… See the full description on the dataset page: https://huggingface.co/datasets/rajpurkar/squad_v2.qa_squadshifts_synthetic
Dataset Card for "lmqg/qa_squadshifts_synthetic"
Dataset Summary
This is a synthetic QA dataset generated with fine-tuned QG models over lmqg/qa_squadshifts, made for question-answering based evaluation (QAE) for question generation model proposed by Zhang and Bansal, 2019.
The test split is the original validation set of lmqg/qa_squadshifts, where the model should be evaluate on.
Supported Tasks and Leaderboards
question-answering
Languages… See the full description on the dataset page: https://huggingface.co/datasets/lmqg/qa_squadshifts_synthetic.squad_kor_v1
Dataset Card for KorQuAD v1.0
Dataset Summary
KorQuAD 1.0 is a large-scale question-and-answer dataset constructed for Korean machine reading comprehension, and investigate the dataset to understand the distribution of answers and the types of reasoning required to answer the question. This dataset benchmarks the data generating process of SQuAD v1.0 to meet the standard.
Supported Tasks and Leaderboards
question-answering
Languages
Korean… See the full description on the dataset page: https://huggingface.co/datasets/KorQuAD/squad_kor_v1.iapp_wiki_qa_squad
iapp_wiki_qa_squad
Extractive question answering over Thai Wikipedia articles, in SQuAD format.
7,242 questions across 1,912 articles, annotated by people iApp hired for the purpose.
from datasets import load_dataset
dataset = load_dataset("iapp/iapp_wiki_qa_squad")
This works again as of the August 2026 revision. Until then it did not. The
repository carried a loading script and no data, and datasets dropped script support
at v3, so load_dataset failed and every… See the full description on the dataset page: https://huggingface.co/datasets/iapp/iapp_wiki_qa_squad.squad_it
Dataset Card for "squad_it"
Dataset Summary
SQuAD-it is derived from the SQuAD dataset and it is obtained through semi-automatic translation of the SQuAD dataset
into Italian. It represents a large-scale dataset for open question answering processes on factoid questions in Italian.
The dataset contains more than 60,000 question/answer pairs derived from the original English dataset. The dataset is
split into training and test sets to support the replicability of the… See the full description on the dataset page: https://huggingface.co/datasets/crux82/squad_it.squad_v1_pt
Dataset Card for "squad_v1_pt"
Dataset Summary
Portuguese translation of the SQuAD dataset. The translation was performed automatically using the Google Cloud API.
Supported Tasks and Leaderboards
More Information Needed
Languages
More Information Needed
Dataset Structure
Data Instances
default
Size of downloaded dataset files: 39.53 MB
Size of the generated dataset: 96.72 MB
Total amount of disk used: 136.25 MB
An… See the full description on the dataset page: https://huggingface.co/datasets/nunorc/squad_v1_pt.Bengali-SQuAD
Overview
This dataset contains the data for the paper Deep learning based question answering system in Bengali. It is a translated version of SQuAD 2.0 dataset to bengali language. Preprocessing details can be found in the paper.
indic-squad
IndicSQuAD Dataset
Dataset Description
IndicSQuAD is a comprehensive multilingual extractive Question Answering (QA) dataset covering nine major Indic languages: Hindi, Bengali, Tamil, Telugu, Marathi, Gujarati, Urdu, Kannada, Oriya, and Malayalam. It's systematically derived from the popular English SQuAD (Stanford Question Answering Dataset).
The rapid progress in QA systems has predominantly benefited high-resource languages, leaving Indic languages significantly… See the full description on the dataset page: https://huggingface.co/datasets/l3cube-pune/indic-squad.squad_id_muis_2020
SQuAD ID
Reorganized and corrected SQuAD ID dataset.
Source
Paper Title: Sequence-to-Sequence Learning for Indonesian Automatic Question Generator
Venue: ICAICTA 2020
Data Correction
I found many incorrect answer_start from the original data, where the extracted subtext from context using the position does not match the answer text. To correct this, a script is used to attempt to find answer text in context using Python string find() method. If found, the… See the full description on the dataset page: https://huggingface.co/datasets/sagaekakristi/squad_id_muis_2020.cs_squad-3.0
Dataset Card for Czech Simple Question Answering Dataset 3.0
This a processed and filtered adaptation of an existing dataset. For raw and larger dataset, see Dataset Source section.
Dataset Description
The data contains questions and answers based on Czech wikipeadia articles.
Each question has an answer (or more) and a selected part of the context as the evidence.
A majority of the answers are extractive - i.e. they are present in the context in the exact form. The… See the full description on the dataset page: https://huggingface.co/datasets/fewshot-goes-multilingual/cs_squad-3.0.TriviaQA_SQuAD
This is a TriviaQA wikipedia dataset that was reformated and "answer_start" added
This dataset has context max tokens length of 5000.
I used this dataset for my research, you can find code for reformatting TriviaQA here:
https://github.com/Kkordik/NovelQSI
spoken_squad
Dataset Card for Spoken-SQuAD
Citation
@article{lee2018spoken,
title={Spoken SQuAD: A Study of Mitigating the Impact of Speech Recognition Errors on Listening Comprehension},
author={Lee, Chia-Hsuan and Wu, Szu-Lin and Liu, Chi-Liang and Lee, Hung-yi},
journal={Proc. Interspeech 2018},
pages={3459--3463},
year={2018}
}
squad-it
Squad-it
This dataset is an adapted version of that squad-it to train on HuggingFace models.
It contains:
train samples: 87599
test samples : 10570
This dataset is for question answering and his format is the following:
[
{
"answers": [
{
"answer_start": [1],
"text": ["Questo è un testo"]
},
],
"context": "Questo è un testo relativo al contesto.",
"id": "1",
"question": "Questo è un testo?",
"title": "train test"
}
]
It can… See the full description on the dataset page: https://huggingface.co/datasets/z-uo/squad-it.finbenchv2-squad-strip-fi-mt
finbenchv2-squad-strip-fi-mt
This dataset is a subset of our SQuAD v2 HF dataset with
unanswerable questions removed, to be used within the FIN-bench-v2 benchmark suite. An additional feature of this
dataset is that the text in the title fields have been machine-translated to Finnish.
Paper: https://huggingface.co/papers/2512.13330
Code: https://github.com/LumiOpen/lm-evaluation-harness
Considerations for Using the Data
Due to DeepL terms and conditions, this… See the full description on the dataset page: https://huggingface.co/datasets/TurkuNLP/finbenchv2-squad-strip-fi-mt.squad_xh
SQuAD isiXhosa
isiXhosa translation of rajpurkar/squad,
built so that the answer is always a verbatim substring of its context — the
property extractive QA depends on.
Why this exists
Translating question, answer and context independently breaks extractive QA:
the translator renders the same phrase differently in each field, so the answer
usually is not present in its own passage. An earlier field-wise translation of
SQuAD-train measured 41.1% answer-in-context —… See the full description on the dataset page: https://huggingface.co/datasets/simpra/squad_xh.squad-nl-v2.0
SQuAD-NL v2.0 [translated SQuAD / XQuAD]
SQuAD-NL v2.0 is a translation of The Stanford Question Answering Dataset (SQuAD) v2.0.
Since the original English SQuAD test data is not public, we reserve the same documents that were used for XQuAD for testing purposes. These documents are sampled from the original dev data split. The English data is automatically translated using Google Translate (February 2023) and the test data is manually post-edited.
This version of SQuAD-NL also… See the full description on the dataset page: https://huggingface.co/datasets/GroNLP/squad-nl-v2.0.chinese-squadv2English | 中文
Dataset Card for Chinese SQuAD 2.0 (revised, bilingual)
Dataset Description
This is a revised and extended version of the Chinese translation of SQuAD 2.0,
originally machine-translated by
ChineseSquad. Like SQuAD 2.0 it
contains both answerable and unanswerable questions and is designed for Chinese
extractive reading comprehension / question answering.
Compared with the previous release of chinese-squadv2, this version:
Adds the original English… See the full description on the dataset page: https://huggingface.co/datasets/real-jiakai/chinese-squadv2.long_squad_v2
Dataset Card for long_squad_v2
long_squad_v2 is a long-context question answering dataset based on the SQuAD v2 format. It was constructed by concatenating multiple SQuAD v2 contexts to significantly increase the average document length, enabling training and evaluation of models on long-range understanding and sparse answer retrieval tasks.
Dataset Details
Uses
To load the dataset using the 🤗 Datasets library:
from datasets import load_dataset
dataset =… See the full description on the dataset page: https://huggingface.co/datasets/huutuan/long_squad_v2.squad-augmented-v2squad-nl-v1.1
SQuAD-NL v1.1 [translated SQuAD / XQuAD]
SQuAD-NL v1.1 is a translation of The Stanford Question Answering Dataset (SQuAD) v1.1.
Since the original English SQuAD test data is not public, we reserve the same documents that were used for XQuAD for testing purposes. These documents are sampled from the original dev data split. The English data is automatically translated using Google Translate (February 2023) and the test data is manually post-edited.
This version of SQuAD-NL only… See the full description on the dataset page: https://huggingface.co/datasets/GroNLP/squad-nl-v1.1.squad-cmrc2018-zhtw
Dataset Card for "squad-cmrc2018-zhtw"
資料集摘要
CMRC 2018 是第二屆「訊飛盃」中文機器閱讀理解頒獎研討會（CMRC 2018）中相關競賽所使用的資料集。
它主要用於中文機器閱讀理解的跨度提取資料集，以增加該領域的語言多樣性。該資料集由人類專家在維基百科段落上註釋的近 20,000 個真實問題組成。
同時它也註釋了一個挑戰集，其中包含需要在整個上下文中進行全面理解和多句推理的問題。
原始資料來源:
https://hfl-rc.github.io/cmrc2018/
https://github.com/ymcui/cmrc2018
資料下載清理
下載 cmrc2018 資料集
使用 OpenCC 來進行簡繁轉換
使用 Python 正規表示式來清理一些殘留在 context, question, answer 的不必要字元
根據 answers.text 來重新計算 answers.answer_start 的字元位置
使用 Huggingface… See the full description on the dataset page: https://huggingface.co/datasets/erhwenkuo/squad-cmrc2018-zhtw.RAGAS_xquad_x_squadtest_half split from XQuAD https://huggingface.co/datasets/xxizhouu/RAGAS_xquad
PLUS
one impossible question(english) for each paragraph, taken from SQuAD 2.0
test_id: shared uuid accross different spilt
cmi: code mix index
squad
Dataset Card for SQuAD
Dataset Summary
Stanford Question Answering Dataset (SQuAD) is a reading comprehension dataset, consisting of questions posed by crowdworkers on a set of Wikipedia articles, where the answer to every question is a segment of text, or span, from the corresponding reading passage, or the question might be unanswerable.
SQuAD 1.1 contains 100,000+ question-answer pairs on 500+ articles.
Supported Tasks and Leaderboards
Question… See the full description on the dataset page: https://huggingface.co/datasets/DAboy12/squad.qasper-squad
Quasper into squad version
This is a change of format of qasper dataset into squad format.
squad_v2.0squad
Dataset Card for SQuAD
Dataset Summary
Stanford Question Answering Dataset (SQuAD) is a reading comprehension dataset, consisting of questions posed by crowdworkers on a set of Wikipedia articles, where the answer to every question is a segment of text, or span, from the corresponding reading passage, or the question might be unanswerable.
SQuAD 1.1 contains 100,000+ question-answer pairs on 500+ articles.
Supported Tasks and Leaderboards
Question… See the full description on the dataset page: https://huggingface.co/datasets/kritikainwati/squad.squad
Dataset Card for SQuAD
Dataset Summary
Stanford Question Answering Dataset (SQuAD) is a reading comprehension dataset, consisting of questions posed by crowdworkers on a set of Wikipedia articles, where the answer to every question is a segment of text, or span, from the corresponding reading passage, or the question might be unanswerable.
SQuAD 1.1 contains 100,000+ question-answer pairs on 500+ articles.
Supported Tasks and Leaderboards
Question Answering.… See the full description on the dataset page: https://huggingface.co/datasets/mahadchangaizkhan/squad.kor_squad_v2
Dataset Card for squad_v2
Licensing Information
The data is distributed under the CC BY SA 4.0 license.
Source Data Citation Information
@article{2016arXiv160605250R,
author = {{Rajpurkar}, Pranav and {Zhang}, Jian and {Lopyrev},
Konstantin and {Liang}, Percy},
title = "{SQuAD: 100,000+ Questions for Machine Comprehension of Text}",
journal = {arXiv e-prints},
year = 2016,
eid =… See the full description on the dataset page: https://huggingface.co/datasets/KETI-NLP/kor_squad_v2.ua-squad
Dataset Card for UAQuAD
This is a revised version of the Ukrainian SQuAD dataset intended for internal use in the HPLT project. The dataset is constructed as follows:
Examples with the answer appearing in the passage more than 1 time are discarded to prevent potential generation of the frequent spans.
Examples with the answer frequency of more than 1 over the dataset are filtered out to prevent potential span frequency bias in the few-shot regimes.
The answer spans are… See the full description on the dataset page: https://huggingface.co/datasets/HPLT/ua-squad.