Omartificial-Intelligence-Space's picture
Update README.md
eb8adf7 verified
metadata
license: cc-by-nc-4.0
gated: true
extra_gated_prompt: Please provide the required information to access this model
extra_gated_fields:
  Full Name:
    type: text
  Affiliation / Company:
    type: text
  Email Address:
    type: text
  Intended Use:
    type: select
    options:
      - Research
      - Education
      - Commercial Exploration
      - Academic Project
      - Other
extra_gated_heading: Access Request โ€“ Provide Required Information
extra_gated_description: Before accessing this model, please complete the form below.
extra_gated_button_content: Submit Access Request
configs:
  - config_name: Communication & Social Media
    data_files:
      - split: train
        path: Communication & Social Media/train-*
  - config_name: Culture & Heritage
    data_files:
      - split: train
        path: Culture & Heritage/train-*
  - config_name: Daily Life & Household
    data_files:
      - split: train
        path: Daily Life & Household/train-*
  - config_name: Education
    data_files:
      - split: train
        path: Education/train-*
  - config_name: Entertainment
    data_files:
      - split: train
        path: Entertainment/train-*
  - config_name: Finance & Banking
    data_files:
      - split: train
        path: Finance & Banking/train-*
  - config_name: Food
    data_files:
      - split: train
        path: Food/train-*
  - config_name: Geography
    data_files:
      - split: train
        path: Geography/train-*
  - config_name: Government Services
    data_files:
      - split: train
        path: Government Services/train-*
  - config_name: History
    data_files:
      - split: train
        path: History/train-*
  - config_name: Medical
    data_files:
      - split: train
        path: Medical/train-*
  - config_name: Nature & Environment
    data_files:
      - split: train
        path: Nature & Environment/train-*
  - config_name: Saudi Anthropology
    data_files:
      - split: train
        path: Saudi Anthropology/train-*
  - config_name: Shopping & Fashion
    data_files:
      - split: train
        path: Shopping & Fashion/train-*
  - config_name: Social Gatherings & Events
    data_files:
      - split: train
        path: Social Gatherings & Events/train-*
  - config_name: Sports & Fitness
    data_files:
      - split: train
        path: Sports & Fitness/train-*
  - config_name: Technology
    data_files:
      - split: train
        path: Technology/train-*
  - config_name: Transportation
    data_files:
      - split: train
        path: Transportation/train-*
  - config_name: Travel
    data_files:
      - split: train
        path: Travel/train-*
  - config_name: Weather & Seasons
    data_files:
      - split: train
        path: Weather & Seasons/train-*
  - config_name: Work & Office
    data_files:
      - split: train
        path: Work & Office/train-*
dataset_info:
  - config_name: Communication & Social Media
    features:
      - name: Anchor
        dtype: string
      - name: Positive
        dtype: string
      - name: Negative
        dtype: string
    splits:
      - name: train
        num_bytes: 26546
        num_examples: 100
    download_size: 17311
    dataset_size: 26546
  - config_name: Culture & Heritage
    features:
      - name: Anchor
        dtype: string
      - name: Positive
        dtype: string
      - name: Negative
        dtype: string
    splits:
      - name: train
        num_bytes: 22676
        num_examples: 102
    download_size: 14135
    dataset_size: 22676
  - config_name: Daily Life & Household
    features:
      - name: Anchor
        dtype: string
      - name: Positive
        dtype: string
      - name: Negative
        dtype: string
    splits:
      - name: train
        num_bytes: 16680
        num_examples: 98
    download_size: 11583
    dataset_size: 16680
  - config_name: Education
    features:
      - name: Anchor
        dtype: string
      - name: Positive
        dtype: string
      - name: Negative
        dtype: string
    splits:
      - name: train
        num_bytes: 36869
        num_examples: 150
    download_size: 23302
    dataset_size: 36869
  - config_name: Entertainment
    features:
      - name: Anchor
        dtype: string
      - name: Positive
        dtype: string
      - name: Negative
        dtype: string
    splits:
      - name: train
        num_bytes: 18468
        num_examples: 106
    download_size: 12344
    dataset_size: 18468
  - config_name: Finance & Banking
    features:
      - name: Anchor
        dtype: string
      - name: Positive
        dtype: string
      - name: Negative
        dtype: string
    splits:
      - name: train
        num_bytes: 51748
        num_examples: 200
    download_size: 13030
    dataset_size: 51748
  - config_name: Food
    features:
      - name: Anchor
        dtype: string
      - name: Positive
        dtype: string
      - name: Negative
        dtype: string
    splits:
      - name: train
        num_bytes: 27939
        num_examples: 200
    download_size: 15511
    dataset_size: 27939
  - config_name: Geography
    features:
      - name: Anchor
        dtype: string
      - name: Positive
        dtype: string
      - name: Negative
        dtype: string
    splits:
      - name: train
        num_bytes: 14147
        num_examples: 91
    download_size: 9469
    dataset_size: 14147
  - config_name: Government Services
    features:
      - name: Anchor
        dtype: string
      - name: Positive
        dtype: string
      - name: Negative
        dtype: string
    splits:
      - name: train
        num_bytes: 38613
        num_examples: 200
    download_size: 21323
    dataset_size: 38613
  - config_name: History
    features:
      - name: Anchor
        dtype: string
      - name: Positive
        dtype: string
      - name: Negative
        dtype: string
    splits:
      - name: train
        num_bytes: 23662
        num_examples: 150
    download_size: 6347
    dataset_size: 23662
  - config_name: Medical
    features:
      - name: Anchor
        dtype: string
      - name: Positive
        dtype: string
      - name: Negative
        dtype: string
    splits:
      - name: train
        num_bytes: 45448
        num_examples: 200
    download_size: 27780
    dataset_size: 45448
  - config_name: Nature & Environment
    features:
      - name: Anchor
        dtype: string
      - name: Positive
        dtype: string
      - name: Negative
        dtype: string
    splits:
      - name: train
        num_bytes: 41532
        num_examples: 200
    download_size: 26052
    dataset_size: 41532
  - config_name: Saudi Anthropology
    features:
      - name: Anchor
        dtype: string
      - name: Positive
        dtype: string
      - name: Negative
        dtype: string
    splits:
      - name: train
        num_bytes: 24738
        num_examples: 104
    download_size: 14684
    dataset_size: 24738
  - config_name: Shopping & Fashion
    features:
      - name: Anchor
        dtype: string
      - name: Positive
        dtype: string
      - name: Negative
        dtype: string
    splits:
      - name: train
        num_bytes: 19514
        num_examples: 100
    download_size: 13309
    dataset_size: 19514
  - config_name: Social Gatherings & Events
    features:
      - name: Anchor
        dtype: string
      - name: Positive
        dtype: string
      - name: Negative
        dtype: string
    splits:
      - name: train
        num_bytes: 17539
        num_examples: 100
    download_size: 12046
    dataset_size: 17539
  - config_name: Sports & Fitness
    features:
      - name: Anchor
        dtype: string
      - name: Positive
        dtype: string
      - name: Negative
        dtype: string
    splits:
      - name: train
        num_bytes: 48747
        num_examples: 200
    download_size: 25461
    dataset_size: 48747
  - config_name: Technology
    features:
      - name: Anchor
        dtype: string
      - name: Positive
        dtype: string
      - name: Negative
        dtype: string
    splits:
      - name: train
        num_bytes: 20857
        num_examples: 100
    download_size: 13317
    dataset_size: 20857
  - config_name: Transportation
    features:
      - name: Anchor
        dtype: string
      - name: Positive
        dtype: string
      - name: Negative
        dtype: string
    splits:
      - name: train
        num_bytes: 13759
        num_examples: 109
    download_size: 10299
    dataset_size: 13759
  - config_name: Travel
    features:
      - name: Anchor
        dtype: string
      - name: Positive
        dtype: string
      - name: Negative
        dtype: string
    splits:
      - name: train
        num_bytes: 26486
        num_examples: 150
    download_size: 13096
    dataset_size: 26486
  - config_name: Weather & Seasons
    features:
      - name: Anchor
        dtype: string
      - name: Positive
        dtype: string
      - name: Negative
        dtype: string
    splits:
      - name: train
        num_bytes: 32327
        num_examples: 200
    download_size: 19256
    dataset_size: 32327
  - config_name: Work & Office
    features:
      - name: Anchor
        dtype: string
      - name: Positive
        dtype: string
      - name: Negative
        dtype: string
    splits:
      - name: train
        num_bytes: 18158
        num_examples: 104
    download_size: 11650
    dataset_size: 18158
language:
  - ar
pretty_name: Saudi Triplet
task_categories:
  - feature-extraction
  - sentence-similarity
tags:
  - saudi
  - Arabic
  - Triplet
size_categories:
  - 1K<n<10K

๐Ÿ“‚ SaudiDialect-Triplet-21 : Saudi Triplet Dataset (SABER Training Data)

๐Ÿงฉ Dataset Summary

The Saudi Triplet Dataset is a high-quality corpus of 2,964 sentence triplets (Anchor, Positive, Negative) specifically curated to capture the nuances of Saudi Arabic dialects (Najdi, Hijazi, Gulf, etc.).

This dataset was created to fine-tune semantic embedding models such as SABER for tasks like Semantic Search, Retrieval-Augmented Generation (RAG), and Clustering.

It covers 21 distinct domains reflecting real-life Saudi contexts, ranging from Government Services and Finance to Tribal Anthropology and Bedouin Culture.

Team

Special thanks to the exceptional team behind this dataset.

Team

โœˆ๏ธ

Travel
Mohammed Alhassan

๐Ÿ”

Food
Abdulelah Alankari

๐Ÿ›๏ธ

Fashion
Reem Alsuliman

๐ŸŽ“

Education
Joud Aloqla

๐Ÿ’ผ

Work
Nouf Alessa

๐Ÿ“ฑ

Tech
Jude Alsubaie

๐Ÿ‹๏ธ

Sports
Albara Aseri

๐Ÿš—

Transport
Wajn Alqahtani

๐ŸŽฌ

Entertainment
Muzon Assiri

๐Ÿ 

Daily Life
Jana Alsuhaibani

๐Ÿ’ฐ

Finance
Abdullah Alsalem

๐ŸŒค๏ธ

Weather
Huda Aldawsari

๐ŸŽ‰

Events
Shaden Alosaimi

๐Ÿฉบ

Medical
Munirah Alsubaie

๐Ÿ“ข

Social
Mohammed Alziyad

๐Ÿ‡ธ๐Ÿ‡ฆ

Culture
Shatha Alotaibi

๐ŸŒฟ

Nature
Norah Altwijri

๐Ÿ“œ

History
Renad Alrifai

๐Ÿ—บ๏ธ

Geography
Murtada Altarouti

๐Ÿ›๏ธ

Gov
Lama Almutairi

๐Ÿ‘ฅ

Anthro
Adnan Hawsawi

๐Ÿ“Š Dataset Statistics

Statistic Value
Total Triplets 2,964
Total Domains 21
Language Saudi Dialect
Duplicate Anchors 59 (Multi-positive/negative pairings)

๐Ÿ“ Sentence Lengths (Word Count)

The dataset consists primarily of short-to-medium length queries and sentences, typical of search and conversational inputs.

Metric Anchor Positive Negative
Mean 6.42 6.50 5.34
Std Dev 1.85 1.96 1.77
Min 2 2 2
Max 13 15 12

๐Ÿ™๏ธ Domain Distribution

The dataset is balanced across high-resource topics (Food, Finance) and specific cultural topics (Anthropology, Heritage).

Domain Count
Food 200
Finance & Banking 200
Government Services 200
Medical 200
Sports & Fitness 200
Weather & Seasons 200
Nature & Environment 200
Education 150
Travel 150
History 150
Transportation 109
Entertainment 106
Saudi Anthropology 104
Work & Office 104
Culture & Heritage 102
Shopping & Fashion 100
Technology 100
Communication & Social Media 100
Social Gatherings & Events 100
Daily Life & Household 98
Geography 91

๐Ÿ“‚ Data Structure

Each row in the dataset represents a training triplet designed for Contrastive Learning (e.g., MNRL).

Column Name Type Description
Anchor String The reference sentence/query in Saudi dialect.
Positive String A sentence semantically similar to the Anchor (paraphrase or answer).
Negative String A sentence semantically dissimilar to the Anchor (different topic or meaning).
Domain String The topic category of the triplet.

๐Ÿ“ Data Samples

Below are real examples from the dataset showing the dialectal variations and domain diversity.

Domain Anchor (Query) Positive (Match) Negative (Mismatch)
Shopping & Fashion ุฃุจูŠ ูุฑุดู‡ ุชููƒ ุงู„ุนู‚ุฏ ูˆู…ุง ุชู‚ุทุน ุงู„ุดุนุฑ ุงุจูŠ ู…ุดุท ู…ุง ูŠุฎุฑุจ ุงู„ุดุนุฑ ูˆูŠู†ุชูู‡ ู…ุชู‰ ุจูŠูˆุตู„ู†ูŠ ุทู‚ู… ุงู„ุฃู„ู…ุงุณ ุงู„ู„ูŠ ุทู„ุจุชู‡ุŸ
Finance & Banking ุฃุจุบุง ุฃูุชุญ ู…ุญูุธุฉ ุฃุณู‡ู… ูˆุฃุจุฏุฃ ุงุณุชุซู…ุงุฑ ุจุณูŠุท ุฃููƒุฑ ุฃุจุฏุฃ ุชุฏุงูˆู„ ุฎููŠู ููŠ ุงู„ุฃุณู‡ู… ุนู† ุทุฑูŠู‚ ุงู„ู…ุญูุธุฉ ู†ุงูˆูŠ ุฃุฒูˆุฑ ุงู„ุนุงุฆู„ุฉ ููŠ ุงู„ู‚ุฑูŠุฉ ุงู„ุฃุณุจูˆุน ุงู„ุฌุงูŠ
Culture & Heritage ุฃู…ุณ ุณู…ุนุช ู‚ุตุงุฆุฏ ุนู† ุงู„ุดุฌุงุนุฉ ูˆุงู„ูุฑูˆุณูŠุฉ ุงู„ู‚ุตุงูŠุฏ ุงู„ุจุฏูˆูŠุฉ ู…ุนุงู†ูŠู‡ุง ู‚ูˆูŠุฉ ุดุบู„ุช ุงู„ุบุณุงู„ุฉ ุจุงู„ุบู„ุท
Food ุงู„ุณูˆูู„ูŠู‡ ุนู†ุฏู‡ู… ูุฎู… ุงู„ุณูˆูู„ูŠู‡ ูŠุฐูˆุจ ุจุงู„ูู… ู…ุง ูˆุตู„ุช ุงู„ุดุญู†ุฉ
History ุงู„ูˆุงู„ุฏ ูƒุงู† ุฏุงูŠู… ูŠุฐูƒุฑ ู…ู…ู„ูƒุฉ ู„ุญูŠุงู† ุดูุช ุจุฑู†ุงู…ุฌ ูŠุชูƒู„ู… ุนู† ุณูˆู‚ ุนูƒุงุธ ุทู„ุจูŠ ุชุฃุฎุฑ ุจุงู„ู…ุทุนู…
Travel ูˆูŠู† ุฃุญุตู„ ุนู„ู‰ ุฌูˆู„ุงุช ุณูŠุงุญูŠุฉ ุฑุฎูŠุตุฉุŸ ุฃุจุบู‰ ุฃู„ู‚ู‰ ุนุฑูˆุถ ุณูŠุงุญูŠุฉ ุงู‚ุชุตุงุฏูŠุฉ ุงู„ุฌูˆ ุญุงุฑ ูˆู…ุง ุฃู‚ุฏุฑ ุฃุทู„ุน

โš ๏ธ Quality & Integrity

  • Missing Data: There are no missing values in the Anchor, Positive, or Negative columns.
  • Duplicates: There are 59 duplicate anchors. This is intentional in some cases to provide multiple positive pairings for the same query or to enforce separation from different hard negatives.
  • Dialect Intensity: The text ranges from "White Dialect" (understandable by most Arabs) to deep Saudi vernacular (specific to Najd/Hijaz/South).

๐Ÿ› ๏ธ Usage

This dataset is optimized for training sentence transformers using MultipleNegativesRankingLoss.

from datasets import load_dataset

# Load the dataset (Example path)
dataset = load_dataset("Omartificial-Intelligence-Space/Saudi-Triplet-Dataset")

# Print first example
print(dataset['train'][0])