Heng666's picture
Update README.md
3208567 verified
metadata
dataset_info:
  - config_name: aya_human_annotated
    features:
      - name: id
        dtype: int64
      - name: inputs
        dtype: string
      - name: targets
        dtype: string
      - name: language
        dtype: string
      - name: script
        dtype: string
    splits:
      - name: test
        num_bytes: 84266
        num_examples: 250
    download_size: 62478
    dataset_size: 84266
  - config_name: dolly_machine_translated
    features:
      - name: id
        dtype: int64
      - name: inputs
        dtype: string
      - name: targets
        dtype: string
      - name: language
        dtype: string
      - name: script
        dtype: string
    splits:
      - name: test
        num_bytes: 299665
        num_examples: 400
    download_size: 211317
    dataset_size: 299665
configs:
  - config_name: aya_human_annotated
    data_files:
      - split: test
        path: aya_human_annotated/test-*
  - config_name: dolly_machine_translated
    data_files:
      - split: test
        path: dolly_machine_translated/test-*
license: apache-2.0
task_categories:
  - question-answering
  - translation
  - summarization
  - zero-shot-classification
language:
  - zh
pretty_name: Heng666/Traditional_Chinese-aya_evaluation_suite
size_categories:
  - 1M<n<10M

Traditional_Chinese_Aya Header

資料集描述

繁體中文 Aya (Traditional Chinese Aya Chinese;TCA):專注於繁體中文處理的 Aya 集合的精選子集

概述

繁體中文 Aya 是一個精心策劃的資料集,源自 CohereForAI 的綜合 Aya 集合,特別關注繁體中文文本資料。 此資料集結合了來自 CohereForAI/aya_evaluation_suite,過濾掉除繁體中文、簡體中文內容之外的所有內容。

目標

繁體中文 Aya 的目標是為研究人員、技術專家和語言學家提供即用型繁體中文文本資源,顯著減少專注於繁體中文的 NLP 和 AI 專案中數據預處理所需的時間和精力。

資料集來源與資訊

  • 資料來源: 從 CohereForAI/aya_evaluation_suite 3 個子集而來。
  • 語言: 繁體中文、簡體中文('zho')
  • 應用: 非常適合語言建模、文本分類、情感分析、和機器翻譯等任務。
  • 論文連結: 2402.06619
  • 維護人: Heng666
  • License: Apache-2.0

使用方法

此資料集是開始繁體中文語言專案(從學術研究到商業應用)的基礎工具。 透過提供預先過濾的繁體中文文本來源,繁體中文 Aya 讓研究人員、技術專家和開發人員能夠直接進行模型訓練、分析和應用程式開發,而無需進行資料清理和語言過濾的初步麻煩。

展示範例

from datasets import load_dataset
dataset = load_dataset("Heng666/Traditional_Chinese-aya_evaluation_suite", "aya_human_annotated")

在上面的程式碼片段中,「aya_dataset」指的是原始 「aya_evaluation_suite」中「aya_human_annotated」子集的繁體中文版本。 您可以透過在載入資料集時指定其名稱來載入其他子集。

訪問和貢獻

可在 Heng666/Traditional_Chinese-aya_evaluation_suite 下的 Hugging Face Hub 上獲取, 繁體中文 Aya 邀請社區做出貢獻。鼓勵用戶提供回饋、提出改進建議。

支持與合作

我們致力於圍繞繁體中文人工智慧和 NLP 研究創造一個包容和支持的環境。如需支援、協作或有關資料集的疑問,請透過 Hugging Face Hub 的討論部分進行聯絡。

Original Dataset Card of Aya by CohereForAI


Aya Header

Dataset Summary

Aya Evaluation Suite contains a total of 26,750 open-ended conversation-style prompts to evaluate multilingual open-ended generation quality.
To strike a balance between language coverage and the quality that comes with human curation, we create an evaluation suite that includes:

  1. human-curated examples in 7 languages (tur, eng, yor, arb, zho, por, tel) → aya-human-annotated.
  2. machine-translations of handpicked examples into 101 languages → dolly-machine-translated.
  3. human-post-edited translations into 6 languages (hin, srp, rus, fra, arb, spa) → dolly-human-edited.

  • Curated by: Contributors of Aya Open Science Intiative, professional annotators, and synthetic generation
  • Language(s): 101 languages
  • License: Apache 2.0
  • Aya Datasets Family:
    Name Explanation
    aya_dataset Human-annotated multilingual instruction finetuning dataset, comprising over 204K instances across 65 languages.
    aya_collection Created by applying instruction-style templates from fluent speakers to 44 datasets, including translations of 19 instruction-style datasets into 101 languages, providing 513M instances for various tasks.
    aya_evaluation_suite A diverse evaluation set for multilingual open-ended generation, featuring 250 culturally grounded prompts in 7 languages, 200 translated prompts in 24 languages, and human-edited versions selected for cross-cultural relevance from English Dolly in 6 languages.

Dataset

The Aya Evaluation Suite includes the following subsets:

  1. aya-human-annotated: 250 original human-written prompts in 7 languages each.
  2. dolly-machine-translated: 200 human-selected prompts from databricks-dolly-15k , automatically translated with the NLLB model from English into 101 languages (114 dialects in total).
  3. dolly-human-edited: 200 dolly-machine-translated prompts post-edited by fluent speakers for 6 languages.

Load with Datasets

To load this dataset consisting of prompt-completions with datasets, you just need to install Datasets as pip install datasets --upgrade and then use the following code:

from datasets import load_dataset

aya_eval = load_dataset("CohereForAI/aya_evaluation_suite", "dataset")

Data Fields

  • id: Unique id of the data point.
  • inputs: Prompt or input to the language model.
  • targets: Completion or output of the language model. (Not applicable for dolly-human-edited)
  • language: The language of the prompt and completion.
  • script: The writing system of the language.
  • source_id: Corresponding original row index from the databricks-dolly-15k dataset (Field applicable only for subsets dolly-machine-translated & dolly-human-edited)

Data Instances

Example data instances from the Aya Evaluation Suite subsets are listed in the toggled sections below.

aya-human-annotated
{
"id": 42,
"inputs": "What day is known as Star Wars Day?",
"targets": "May 4th (May the 4th be with you!)",
"language": "eng",
"script": "Latn",
}

Dolly-machine-translated and dolly-human-edited

  • These two subsets are parallel datasets (data instances can be mapped using their id column).
  • Note that in the dolly-machine-translated subset, we also include the original English subset (id 1-200), which is translated into 101 languages. Furthermore, the field id can be used to match the translations of the same data instance across languages.
  • The source_id field contains the corresponding original row index from the databricks-dolly-15k dataset.
    dolly-machine-translated
    {
    "id": 2,
    "inputs": "How to escape from a helicopter trapped in water ?",
    "targets": "If you are ever trapped inside a helicopter while submerged in water, it’s best to try and remain calm until the cabin is completely underwater. It’s better to wait for pressure to be equalized, before you try to open the door or break the glass to escape.",
    "language": "eng",
    "script": "Latn",
    "source_id": 6060,
    }
    
    dolly-human-edited
    {
    "id": 2,
    "inputs": "Comment peut-on s'échapper d'un hélicoptère piégé dans l'eau ?",
    "targets": "-",
    "language": "fra",
    "script": "Latn",
    "source_id": 6060,
    }
    

Statistics

The toggled table below lists the breakdown of languages in each subset.

Languages

aya-human-annotated
ISO Code Language Resources
tel Telugu Low
yor Yorùbá Low
arb Arabic High
tur Turkish High
por Portuguese High
zho Chinese (Simplified) High
eng English High
dolly-machine-translated
ISO Code Language Resources
ace Achinese Low
afr Afrikaans Mid
amh Amharic Low
ara (arb, acm, acq, aeb, ajp, apc, ars, ary & arz) Arabic (Standard, Gelet Iraqi, Ta'izzi-Adeni, Tunisian, South Levantine, North Levantine, Najdi, Moroccan & Egyptian) High
aze (azb & azj) Azerbaijani (South & North) Low
bel Belarusian Mid
ben Bengali Mid
bjn Banjar Low
bul Bulgarian Mid
cat Catalan High
ceb Cebuano Mid
ces Czech High
cym Welsh Low
dan Danish Mid
deu German High
ell Greek Mid
eng English High
epo Esperanto Low
est Estonian Mid
eus Basque High
fin Finnish High
fra French High
gla Scottish Gaelic Low
gle Irish Low
glg Galician Mid
guj Gujarati Low
hat Haitian Creole Low
hau Hausa Low
heb Hebrew Mid
hin Hindi High
hun Hungarian High
hye Armenian Low
ibo Igbo Low
ind Indonesian Mid
isl Icelandic Low
ita Italian High
jav Javanese Low
jpn Japanese High
kan Kannada Low
kas Kashmiri Low
kat Georgian Mid
kau (knc) Kanuri (Central) Low
kaz Kazakh Mid
khm Khmer Low
kir Kyrgyz Low
kor Korean High
kur (ckb & kmr) Kurdish (Central & Northern) Low
lao Lao Low
lav (lvs) Latvian (Standard) Mid
lit Lithuanian Mid
ltz Luxembourgish Low
mal Malayalam Low
mar Marathi Low
min Minangkabau Low
mkd Macedonian Low
mlg (plt) Malagasy (Plateau) Low
mlt Maltese Low
mni Manipuri Low
mon (khk) Mongolian (Khalkha) Low
mri Maori Low
msa (zsm) Malay (Standard) Mid
mya Burmese Low
nep (npi) Nepali Low
nld Dutch High
nor (nno & nob) Norwegian (Nynorsk & Bokmål) Low
nso Northern Sotho Low
pes Persian High
pol Polish High
por Portuguese High
pus (pbt) Pashto (Southern) Low
ron Romanian Mid
rus Russian High
sin Sinhala Low
slk Slovak Mid
slv Slovenian Mid
smo Samoan Low
sna Shona Low
snd Sindhi Low
som Somali Low
sot Southern Sotho Low
spa Spanish High
sqi (als) Albanian (Tosk) Low
srp Serbian High
sun Sundanese Low
swa (swh) Swahili (Coastal) Low
swe Swedish High
tam Tamil Mid
taq Tamasheq Low
tel Telugu Low
tgk Tajik Low
tha Thai Mid
tur Turkish High
ukr Ukrainian Mid
urd Urdu Mid
uzb (uzn) Uzbek (Nothern) Mid
vie Vietnamese High
xho Xhosa Low
yid (ydd) Yiddish (Eastern) Low
yor Yoruba Low
zho (+ yue) Chinese (Simplified & Cantonese) High
zul Zulu Low
dolly-human-edited
ISO Code Language Resources
arb Arabic High
fra French High
hin Hindi High
rus Russian High
spa Spanish High
srp Serbian High

Motivations & Intentions

  • Curation Rationale: This evaluation suite is tailored to test the generation quality of multilingual models, with the aim of balancing language coverage and human-sourced quality. It covers prompts originally written in each language, as well as English-centric translated, and manually curated or edited prompts for a linguistically broad, but rich testbed. The list of languages was initially established from mT5 and aligned with the annotators’ language list and the NLLB translation model.

Known Limitations

  • Translation Quality: Note that the expressiveness of the dolly-machine-translated subset is limited by the quality of the translation model and may adversely impact an estimate of ability in languages where translations are not adequate. If this subset is used for testing, we recommend it be paired and reported with the professionally post-edited dolly-human-edited subset or the aya-human-annotated set, which, while covering only 7 languages, is entirely created by proficient target language speakers.

Additional Information

Provenance

  • Methods Used: combination of original annotations by volunteers, automatic translation, and post-editing of translations by professional annotators.
  • Methodology Details:
    • Source: Original annotations from Aya dataset along with translations and post-edits of Dolly dataset
    • Platform: Aya Annotation Platform
    • Dates of Collection: May 2023 - Dec 2023

Dataset Version and Maintenance

  • Maintenance Status: Actively Maintained
  • Version Details:
    • Current version: 1.0
    • Last Update: 02/2024
    • First Release: 02/2024
  • Maintenance Plan: No updates planned.

Authorship

Licensing Information

This dataset can be used for any purpose, whether academic or commercial, under the terms of the Apache 2.0 License.

Citation Information

@misc{singh2024aya,
      title={Aya Dataset: An Open-Access Collection for Multilingual Instruction Tuning}, 
      author={Shivalika Singh and Freddie Vargus and Daniel Dsouza and Börje F. Karlsson and Abinaya Mahendiran and Wei-Yin Ko and Herumb Shandilya and Jay Patel and Deividas Mataciunas and Laura OMahony and Mike Zhang and Ramith Hettiarachchi and Joseph Wilson and Marina Machado and Luisa Souza Moura and Dominik Krzemiński and Hakimeh Fadaei and Irem Ergün and Ifeoma Okoh and Aisha Alaagib and Oshan Mudannayake and Zaid Alyafeai and Vu Minh Chien and Sebastian Ruder and Surya Guthikonda and Emad A. Alghamdi and Sebastian Gehrmann and Niklas Muennighoff and Max Bartolo and Julia Kreutzer and Ahmet Üstün and Marzieh Fadaee and Sara Hooker},
      year={2024},
      eprint={2402.06619},
      archivePrefix={arXiv},
      primaryClass={cs.CL}
}