The dataset viewer should be available soon. Please retry later.
google/xtreme(複製)
データの内容
40 の類型的に多様な言語で言語横断の汎化を測る XTREME ベンチマークのデータを、タスクと言語の組ごとにまとめたもの。構成名は MLQA.ar.ar のように付けられ、列には tokens(単語の列)と ner_tags(タグの列)がある。非商用の部分を含むと調査メモにある。
データの概要
| 項目 | 値 |
|---|---|
| 行数 | 2,768,692(183 構成・392 分割の合計) |
| 列数 | 2〜5(構成による) |
| 容量 | 363.56 MB(ファイル 394 本) |
| ファイルの形式 | parquet 199 |
| 言語 | af, ar, bg, bn, de, el, en, es, et, eu, fa, fi, fr, he, hi, hu, id, it, ja, jv, ka, kk, ko, ml, mr, ms, my, nl, pt, ru, |
| 公開 | public |
構成と分割
| 構成(config) | 分割(split)ごとの行数 | 列数 |
|---|---|---|
MLQA.ar.ar |
test 5,335、validation 517 |
5 |
MLQA.ar.de |
test 1,649、validation 207 |
5 |
MLQA.ar.en |
test 5,335、validation 517 |
5 |
MLQA.ar.es |
test 1,978、validation 161 |
5 |
MLQA.ar.hi |
test 1,831、validation 186 |
5 |
MLQA.ar.vi |
test 2,047、validation 163 |
5 |
MLQA.ar.zh |
test 1,912、validation 188 |
5 |
MLQA.de.ar |
test 1,649、validation 207 |
5 |
MLQA.de.de |
test 4,517、validation 512 |
5 |
MLQA.de.en |
test 4,517、validation 512 |
5 |
MLQA.de.es |
test 1,776、validation 196 |
5 |
MLQA.de.hi |
test 1,430、validation 163 |
5 |
MLQA.de.vi |
test 1,675、validation 182 |
5 |
MLQA.de.zh |
test 1,621、validation 190 |
5 |
MLQA.en.ar |
test 5,335、validation 517 |
5 |
MLQA.en.de |
test 4,517、validation 512 |
5 |
MLQA.en.en |
test 11,590、validation 1,148 |
5 |
MLQA.en.es |
test 5,253、validation 500 |
5 |
MLQA.en.hi |
test 4,918、validation 507 |
5 |
MLQA.en.vi |
test 5,495、validation 511 |
5 |
(ほか 163 構成。全体は HF のデータセットビューアで見られる)
列(構成 PAN-X.af・分割 train。構成ごとに列が違う)
| 列 | 型 | 先頭の行の値 |
|---|---|---|
tokens |
list<string> |
[14 件] |
ner_tags |
list<class_label(7: O, B-PER, I-PER, B-ORG, I-ORG, B-LOC …)> |
[14 件] |
langs |
list<string> |
[14 件] |
行数と列は HF のデータセットビューアの集計(元のデータセットの現在の版で数えたもの)。全データセットの一覧は HYUKI5/dataset-catalog にある。
複製について
google/xtreme をリビジョン ec5f1f46e9af79639a90684a7a70a956c4998f04 の時点でそのまま複製したもの。中身は変えていない。判断モデル Laya(convaiinnovations/laya)の微調整に使う候補として、調査メモ(Laya 学習用データセット調査)に載せたデータを集めている。
出典
- 元のデータセット: https://huggingface.co/datasets/google/xtreme
- 複製したリビジョン:
ec5f1f46e9af79639a90684a7a70a956c4998f04(2024-02-22 17:12:06+00:00) - 元のライセンス表記:
['apache-2.0', 'cc-by-4.0', 'cc-by-2.0', 'cc-by-sa-4.0', 'other', 'cc-by-nc-4.0']
論文
アノテーションの方法
キュレーションの根拠
初期のデータ収集と正規化
元データの言語の作成者は誰か?
アノテーションの過程
微調整に使うときの注意
- 調査メモの台帳での区分: C(NC・ND・研究専用・評価専用など、商用の利用を妨げうる記載がある)。区分は記載の文言で機械的に付けたもので、法的な判定ではない。学習に入れる前に元のカードとライセンスの原文で確かめる。
- Laya の入力は 2,048 token まで(ModernBERT-large / mmBERT)。長い本文は先頭から黙って切られるので、切れる行は事前に除くか、要約・切り出しをしてから入れる。
- 分割は記事・文書単位で行い、同じ本文から作った別の問いを学習側と較正側に分けない。温度と閾値は学習に使っていない分割で別々に当てる。
調査メモの記述
| メモ | 節 | ライセンス(記載) | 確認 | 注記 |
|---|---|---|---|---|
multilingual_cjk.md |
1. 大規模多言語データセット(XNLI, PAWS-X, MASSIVE, SIB-200, Belebele ほか) (L54) | 構成ごとに混在: apache-2.0, cc-by-4.0, cc-by-2.0, cc-by-sa-4.0, other, cc-by-nc-4.0 | 確認済 | 複数/混在 / 確認印なし(各ファイルの凡例により確認済扱い) / 原文注記: dl=31,709。非商用部分あり |
元のカード(日本語訳)
以下は元のデータセットのカードの本文を日本語に訳したもの(機械翻訳)。訳に疑問があるときは、末尾の原文を確かめる。
"xtreme" のデータセットカード
目次
- データセットの説明
- データセットの概要
- 対応するタスクとリーダーボード
- 言語
- データセットの構造
- データのインスタンス
- データのフィールド
- データ分割
- データセットの作成
- キュレーションの根拠
- 元データ
- アノテーション
- 個人情報と機微な情報
- データ利用上の留意点
- データセットの社会的影響
- バイアスに関する検討
- その他の既知の制約
- 補足情報
- データセットのキュレーター
- ライセンス情報
- 引用情報
- 貢献
データセットの説明
- ホームページ: https://github.com/google-research/xtreme
- リポジトリ: More Information Needed
- 論文: More Information Needed
- 連絡先: More Information Needed
- ダウンロードしたデータセットファイルのサイズ: 15.88 GB
- 生成されたデータセットのサイズ: 1.08 GB
- ディスク使用量の合計: 16.96 GB
データセットの概要
Cross-lingual Natural Language Inference(XNLI)コーパスは、MultiNLI コーパスのために、クラウドソーシングで集められた 5,000 組のテスト用ペアと 2,500 組の開発用ペアからなる。各ペアにはテキスト含意(textual entailment)が付与され、フランス語、スペイン語、ドイツ語、ギリシャ語、ブルガリア語、ロシア語、トルコ語、アラビア語、ベトナム語、タイ語、中国語、ヒンディー語、スワヒリ語、ウルドゥー語の 14 言語に翻訳されている。これにより 112.5k の注釈付きペアが得られる。各前提文は、15 言語それぞれの対応する仮説文と結び付けられ、組み合わせは合計 150 万を超える。このコーパスは、訓練時に英語の NLI データしか利用できない場合に、任意の言語(スワヒリ語やウルドゥー語のような低リソース言語を含む)で推論をどのように行うかを評価するために作られている。解決策の一つは言語横断の文エンコーディングであり、XNLI はそれを評価するベンチマークである。
Cross-lingual TRansfer Evaluation of Multilingual Encoders(XTREME)ベンチマークは、事前学習済みの多言語モデルについて、言語横断の汎化能力を評価するためのベンチマークである。40 の類型的に多様な言語(12 の語族にまたがる)を対象とし、構文と意味の異なる水準の推論をまとめて要求する 9 つのタスクを含む。XTREME の言語は、言語の多様性、既存タスクでのカバー範囲、訓練データの入手しやすさを最大化するように選ばれている。その中には研究の少ない言語も多く、たとえばドラヴィダ語族のタミル語(インド南部、スリランカ、シンガポールで話される)、テルグ語とマラヤーラム語(主にインド南部で話される)、ニジェール・コンゴ語族のスワヒリ語とヨルバ語(アフリカで話される)などがある。
対応するタスクとリーダーボード
言語
データセットの構造
データのインスタンス
MLQA.ar.ar
- ダウンロードしたデータセットファイルのサイズ: 75.72 MB
- 生成されたデータセットのサイズ: 9.20 MB
- ディスク使用量の合計: 84.91 MB
'validation' の例は次のとおり。
MLQA.ar.de
- ダウンロードしたデータセットファイルのサイズ: 75.72 MB
- 生成されたデータセットのサイズ: 2.55 MB
- ディスク使用量の合計: 78.27 MB
'validation' の例は次のとおり。
MLQA.ar.en
- ダウンロードしたデータセットファイルのサイズ: 75.72 MB
- 生成されたデータセットのサイズ: 9.04 MB
- ディスク使用量の合計: 84.76 MB
'validation' の例は次のとおり。
MLQA.ar.es
- ダウンロードしたデータセットファイルのサイズ: 75.72 MB
- 生成されたデータセットのサイズ: 3.27 MB
- ディスク使用量の合計: 78.99 MB
'validation' の例は次のとおり。
MLQA.ar.hi
- ダウンロードしたデータセットファイルのサイズ: 75.72 MB
- 生成されたデータセットのサイズ: 3.32 MB
- ディスク使用量の合計: 79.04 MB
'validation' の例は次のとおり。
データのフィールド
すべての分割でフィールドは共通である。
MLQA.ar.ar
id:string型の feature。title:string型の feature。context:string型の feature。question:string型の feature。answers: 以下を含む dictionary 型の feature:answer_start:int32型の feature。text:string型の feature。
MLQA.ar.de
id:string型の feature。title:string型の feature。context:string型の feature。question:string型の feature。answers: 以下を含む dictionary 型の feature:answer_start:int32型の feature。text:string型の feature。
MLQA.ar.en
id:string型の feature。title:string型の feature。context:string型の feature。question:string型の feature。answers: 以下を含む dictionary 型の feature:answer_start:int32型の feature。text:string型の feature。
MLQA.ar.es
id:string型の feature。title:string型の feature。context:string型の feature。question:string型の feature。answers: 以下を含む dictionary 型の feature:answer_start:int32型の feature。text:string型の feature。
MLQA.ar.hi
id:string型の feature。title:string型の feature。context:string型の feature。question:string型の feature。answers: 以下を含む dictionary 型の feature:answer_start:int32型の feature。text:string型の feature。
データ分割
| name | validation | test |
|---|---|---|
| MLQA.ar.ar | 517 | 5335 |
| MLQA.ar.de | 207 | 1649 |
| MLQA.ar.en | 517 | 5335 |
| MLQA.ar.es | 161 | 1978 |
| MLQA.ar.hi | 186 | 1831 |
データセットの作成
キュレーションの根拠
元データ
初期のデータ収集と正規化
元データの言語の作成者は誰か?
アノテーション
アノテーションの過程
アノテーターは誰か?
個人情報と機微な情報
データ利用上の留意点
データセットの社会的影響
バイアスに関する検討
その他の既知の制約
補足情報
データセットのキュレーター
ライセンス情報
引用情報
@InProceedings{conneau2018xnli,
author = {Conneau, Alexis
and Rinott, Ruty
and Lample, Guillaume
and Williams, Adina
and Bowman, Samuel R.
and Schwenk, Holger
and Stoyanov, Veselin},
title = {XNLI: Evaluating Cross-lingual Sentence Representations},
booktitle = {Proceedings of the 2018 Conference on Empirical Methods
in Natural Language Processing},
year = {2018},
publisher = {Association for Computational Linguistics},
location = {Brussels, Belgium},
}
@article{hu2020xtreme,
author = {Junjie Hu and Sebastian Ruder and Aditya Siddhant and Graham Neubig and Orhan Firat and Melvin Johnson},
title = {XTREME: A Massively Multilingual Multi-task Benchmark for Evaluating Cross-lingual Generalization},
journal = {CoRR},
volume = {abs/2003.11080},
year = {2020},
archivePrefix = {arXiv},
eprint = {2003.11080}
}
貢献
このデータセットを追加してくれた @thomwolf、@jplu、@lewtun、@lvwerra、@lhoestq、@patrickvonplaten、@mariamabarham に感謝する。
原文(英語)
Dataset Card for "xtreme"
Dataset Summary
The Cross-lingual Natural Language Inference (XNLI) corpus is a crowd-sourced collection of 5,000 test and 2,500 dev pairs for the MultiNLI corpus. The pairs are annotated with textual entailment and translated into 14 languages: French, Spanish, German, Greek, Bulgarian, Russian, Turkish, Arabic, Vietnamese, Thai, Chinese, Hindi, Swahili and Urdu. This results in 112.5k annotated pairs. Each premise can be associated with the corresponding hypothesis in the 15 languages, summing up to more than 1.5M combinations. The corpus is made to evaluate how to perform inference in any language (including low-resources ones like Swahili or Urdu) when only English NLI data is available at training time. One solution is cross-lingual sentence encoding, for which XNLI is an evaluation benchmark. The Cross-lingual TRansfer Evaluation of Multilingual Encoders (XTREME) benchmark is a benchmark for the evaluation of the cross-lingual generalization ability of pre-trained multilingual models. It covers 40 typologically diverse languages (spanning 12 language families) and includes nine tasks that collectively require reasoning about different levels of syntax and semantics. The languages in XTREME are selected to maximize language diversity, coverage in existing tasks, and availability of training data. Among these are many under-studied languages, such as the Dravidian languages Tamil (spoken in southern India, Sri Lanka, and Singapore), Telugu and Malayalam (spoken mainly in southern India), and the Niger-Congo languages Swahili and Yoruba, spoken in Africa.
Supported Tasks and Leaderboards
Languages
Dataset Structure
Data Instances
MLQA.ar.ar
- Size of downloaded dataset files: 75.72 MB
- Size of the generated dataset: 9.20 MB
- Total amount of disk used: 84.91 MB
An example of 'validation' looks as follows.
MLQA.ar.de
- Size of downloaded dataset files: 75.72 MB
- Size of the generated dataset: 2.55 MB
- Total amount of disk used: 78.27 MB
An example of 'validation' looks as follows.
MLQA.ar.en
- Size of downloaded dataset files: 75.72 MB
- Size of the generated dataset: 9.04 MB
- Total amount of disk used: 84.76 MB
An example of 'validation' looks as follows.
MLQA.ar.es
- Size of downloaded dataset files: 75.72 MB
- Size of the generated dataset: 3.27 MB
- Total amount of disk used: 78.99 MB
An example of 'validation' looks as follows.
MLQA.ar.hi
- Size of downloaded dataset files: 75.72 MB
- Size of the generated dataset: 3.32 MB
- Total amount of disk used: 79.04 MB
An example of 'validation' looks as follows.
Data Fields
The data fields are the same among all splits.
MLQA.ar.ar
id: astringfeature.title: astringfeature.context: astringfeature.question: astringfeature.answers: a dictionary feature containing:answer_start: aint32feature.text: astringfeature.
MLQA.ar.de
id: astringfeature.title: astringfeature.context: astringfeature.question: astringfeature.answers: a dictionary feature containing:answer_start: aint32feature.text: astringfeature.
MLQA.ar.en
id: astringfeature.title: astringfeature.context: astringfeature.question: astringfeature.answers: a dictionary feature containing:answer_start: aint32feature.text: astringfeature.
MLQA.ar.es
id: astringfeature.title: astringfeature.context: astringfeature.question: astringfeature.answers: a dictionary feature containing:answer_start: aint32feature.text: astringfeature.
MLQA.ar.hi
id: astringfeature.title: astringfeature.context: astringfeature.question: astringfeature.answers: a dictionary feature containing:answer_start: aint32feature.text: astringfeature.
Data Splits
| name | validation | test |
|---|---|---|
| MLQA.ar.ar | 517 | 5335 |
| MLQA.ar.de | 207 | 1649 |
| MLQA.ar.en | 517 | 5335 |
| MLQA.ar.es | 161 | 1978 |
| MLQA.ar.hi | 186 | 1831 |
Dataset Creation
Curation Rationale
Source Data
Initial Data Collection and Normalization
Who are the source language producers?
Annotations
Annotation process
Who are the annotators?
Personal and Sensitive Information
Considerations for Using the Data
Social Impact of Dataset
Discussion of Biases
Other Known Limitations
Additional Information
Dataset Curators
Licensing Information
Citation Information
@InProceedings{conneau2018xnli,
author = {Conneau, Alexis
and Rinott, Ruty
and Lample, Guillaume
and Williams, Adina
and Bowman, Samuel R.
and Schwenk, Holger
and Stoyanov, Veselin},
title = {XNLI: Evaluating Cross-lingual Sentence Representations},
booktitle = {Proceedings of the 2018 Conference on Empirical Methods
in Natural Language Processing},
year = {2018},
publisher = {Association for Computational Linguistics},
location = {Brussels, Belgium},
}
@article{hu2020xtreme,
author = {Junjie Hu and Sebastian Ruder and Aditya Siddhant and Graham Neubig and Orhan Firat and Melvin Johnson},
title = {XTREME: A Massively Multilingual Multi-task Benchmark for Evaluating Cross-lingual Generalization},
journal = {CoRR},
volume = {abs/2003.11080},
year = {2020},
archivePrefix = {arXiv},
eprint = {2003.11080}
}
Contributions
Thanks to @thomwolf, @jplu, @lewtun, @lvwerra, @lhoestq, @patrickvonplaten, @mariamabarham for adding this dataset.
- Downloads last month
- -