README.md

metadata

dataset_info:
  features:
    - name: split
      dtype: string
    - name: score
      dtype: float64
    - name: sentence1
      dtype: string
    - name: sentence2
      dtype: string
  splits:
    - name: train
      num_bytes: 1222113
      num_examples: 5749
    - name: validation
      num_bytes: 334413
      num_examples: 1500
    - name: test
      num_bytes: 264980
      num_examples: 1379
  download_size: 917266
  dataset_size: 1821506
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
      - split: validation
        path: data/validation-*
      - split: test
        path: data/test-*
language:
  - ar

Arabic Translation of STS Benchmark Dataset

This repository contains an Arabic translation of the "mteb/stsbenchmark-sts" dataset . The original dataset is designed for evaluating sentence similarity models, and this translated version enables similar research and development in the Arabic language.

Translation Details

The translation was performed using the "Helsinki-NLP/opus-mt-en-ar" model from Hugging Face.

How to Use

from datasets import load_dataset

dataset = load_dataset("Ruqiya/Arabic-stsbenchmark-sts-ar")