Hugging Face's logo

Datasets:
agentlans
/
multilingual-sentences

Tasks:
Text Generation
Text Classification
Text Retrieval
Modalities:
Text
Formats:
text
Languages:
multilingual
Arabic
Azerbaijani
Size:
10M - 100M
Libraries:
Datasets
Dataset card Data Studio Files Community
2
multilingual-sentences / README.md
agentlans's picture
agentlans
Update README.md
9992209 verified
preview code
|
raw
history blame contribute delete
7.4 kB
metadata
configs:
  - config_name: all
    data_files:
      - path:
          - all.txt.zst
        split: train
    default: true
  - config_name: ar
    data_files:
      - path:
          - ar.txt.zst
        split: train
  - config_name: az
    data_files:
      - path:
          - az.txt.zst
        split: train
  - config_name: bg
    data_files:
      - path:
          - bg.txt.zst
        split: train
  - config_name: bn
    data_files:
      - path:
          - bn.txt.zst
        split: train
  - config_name: ca
    data_files:
      - path:
          - ca.txt.zst
        split: train
  - config_name: cs
    data_files:
      - path:
          - cs.txt.zst
        split: train
  - config_name: da
    data_files:
      - path:
          - da.txt.zst
        split: train
  - config_name: de
    data_files:
      - path:
          - de.txt.zst
        split: train
  - config_name: el
    data_files:
      - path:
          - el.txt.zst
        split: train
  - config_name: en
    data_files:
      - path:
          - en.txt.zst
        split: train
  - config_name: es
    data_files:
      - path:
          - es.txt.zst
        split: train
  - config_name: et
    data_files:
      - path:
          - et.txt.zst
        split: train
  - config_name: fa
    data_files:
      - path:
          - fa.txt.zst
        split: train
  - config_name: fi
    data_files:
      - path:
          - fi.txt.zst
        split: train
  - config_name: fr
    data_files:
      - path:
          - fr.txt.zst
        split: train
  - config_name: he
    data_files:
      - path:
          - he.txt.zst
        split: train
  - config_name: hi
    data_files:
      - path:
          - hi.txt.zst
        split: train
  - config_name: hu
    data_files:
      - path:
          - hu.txt.zst
        split: train
  - config_name: hy
    data_files:
      - path:
          - hy.txt.zst
        split: train
  - config_name: id
    data_files:
      - path:
          - id.txt.zst
        split: train
  - config_name: is
    data_files:
      - path:
          - is.txt.zst
        split: train
  - config_name: it
    data_files:
      - path:
          - it.txt.zst
        split: train
  - config_name: ja
    data_files:
      - path:
          - ja.txt.zst
        split: train
  - config_name: ka
    data_files:
      - path:
          - ka.txt.zst
        split: train
  - config_name: kk
    data_files:
      - path:
          - kk.txt.zst
        split: train
  - config_name: ko
    data_files:
      - path:
          - ko.txt.zst
        split: train
  - config_name: lt
    data_files:
      - path:
          - lt.txt.zst
        split: train
  - config_name: lv
    data_files:
      - path:
          - lv.txt.zst
        split: train
  - config_name: mk
    data_files:
      - path:
          - mk.txt.zst
        split: train
  - config_name: ml
    data_files:
      - path:
          - ml.txt.zst
        split: train
  - config_name: mr
    data_files:
      - path:
          - mr.txt.zst
        split: train
  - config_name: ne
    data_files:
      - path:
          - ne.txt.zst
        split: train
  - config_name: nl
    data_files:
      - path:
          - nl.txt.zst
        split: train
  - config_name: 'no'
    data_files:
      - path:
          - no.txt.zst
        split: train
  - config_name: pl
    data_files:
      - path:
          - pl.txt.zst
        split: train
  - config_name: pt
    data_files:
      - path:
          - pt.txt.zst
        split: train
  - config_name: ro
    data_files:
      - path:
          - ro.txt.zst
        split: train
  - config_name: ru
    data_files:
      - path:
          - ru.txt.zst
        split: train
  - config_name: sk
    data_files:
      - path:
          - sk.txt.zst
        split: train
  - config_name: sl
    data_files:
      - path:
          - sl.txt.zst
        split: train
  - config_name: sq
    data_files:
      - path:
          - sq.txt.zst
        split: train
  - config_name: sr
    data_files:
      - path:
          - sr.txt.zst
        split: train
  - config_name: sv
    data_files:
      - path:
          - sv.txt.zst
        split: train
  - config_name: ta
    data_files:
      - path:
          - ta.txt.zst
        split: train
  - config_name: th
    data_files:
      - path:
          - th.txt.zst
        split: train
  - config_name: tr
    data_files:
      - path:
          - tr.txt.zst
        split: train
  - config_name: uk
    data_files:
      - path:
          - uk.txt.zst
        split: train
  - config_name: ur
    data_files:
      - path:
          - ur.txt.zst
        split: train
  - config_name: vi
    data_files:
      - path:
          - vi.txt.zst
        split: train
  - config_name: zh
    data_files:
      - path:
          - zh.txt.zst
        split: train
language:
  - multilingual
  - ar
  - az
  - bg
  - bn
  - ca
  - cs
  - da
  - de
  - el
  - en
  - es
  - et
  - fa
  - fi
  - fr
  - he
  - hi
  - hu
  - hy
  - id
  - is
  - it
  - ja
  - ka
  - kk
  - ko
  - lt
  - lv
  - mk
  - ml
  - mr
  - ne
  - nl
  - 'no'
  - pl
  - pt
  - ro
  - ru
  - sk
  - sl
  - sq
  - sr
  - sv
  - ta
  - th
  - tr
  - uk
  - ur
  - vi
  - zh
task_categories:
  - text-generation
  - text-classification
  - text-retrieval
size_categories:
  - 1M<n<10M

Multilingual Sentences

Dataset contains sentences from 50 languages, grouped by their two-letter ISO 639-1 codes. The "all" configuration includes sentences from all languages.

Dataset Overview

Multilingual Sentence Dataset is a comprehensive collection of high-quality, linguistically diverse sentences. Dataset is designed to support a wide range of natural language processing tasks, including but not limited to language modeling, machine translation, and cross-lingual studies.

Methods

Rigorous methodology consisted of three main stages: text preprocessing, language detection, and dataset processing.

Text Preprocessing

Sophisticated text cleaning pipeline using the textacy library, which included:

  • Removal of HTML tags, email addresses, URLs, and emojis
  • Unicode and whitespace normalization
  • Standardization of punctuation and word formats

Language Detection

Google CLD3 library utilized for accurate language identification:

  • Implemented NNetLanguageIdentifier
  • Configured for processing texts between 0-1000 bytes
  • Included reliability assessment for each language detection

Dataset Processing

Workflow for dataset creation involved the following steps:

  1. Streamed loading of the LinguaNova multilingual dataset
  2. Application of the text preprocessing pipeline
  3. Sentence segmentation using PyICU for accurate boundary detection
  4. Quality filtering:
    • Length constraint (maximum 2048 characters per sentence)
    • High-reliability language verification
  5. Extraction of unique sentences
  6. Random shuffling for unbiased sampling
  7. Generation of language-specific files

Technical Details

Libraries and Tools

  • textacy: Advanced text preprocessing
  • Google CLD3: State-of-the-art language detection
  • Hugging Face datasets: Efficient data handling and processing
  • SentenceBreaker: Accurate sentence segmentation

Implementation Notes

  • Process was executed consistently across all 50 languages to ensure uniformity and high quality in the multilingual dataset preparation.
  • Special attention was given to maintaining the integrity of each language's unique characteristics throughout the processing pipeline.

Data Splits

Dataset is organized into the following splits:

  • Individual language files: Contains sentences for each of the 50 languages
  • "all" configuration: Aggregates sentences from all languages into a single dataset

Limitations and Biases

While extensive efforts were made to ensure dataset quality, users should be aware of potential limitations:

  • Language detection accuracy may vary for very short texts or closely related languages
  • Dataset may not fully represent all dialects or regional variations within each language
  • Potential biases in the original LinguaNova dataset could be carried over

Ethical Considerations

Users of this dataset should be mindful of:

  • Potential biases in language representation
  • Need for responsible use in AI applications, especially in multilingual contexts
  • Privacy considerations, although personal identifiable information should have been removed