CoolFace
Datasetpublic

SAWithanage/en-si-parallel-3k

Dataset Card for en-si-parallel-3k Dataset Summary The en-si-parallel-3k dataset is a high-quality, synthetically generated parallel corpus containing 3,000 English-Sinhala translation pairs. It is specifically designed for fine-tuning Large Language Models (LLMs) to enhance English-to-Sinhala translation capabilities and cross-lingual understanding. Dataset Composition The dataset is structured into 60 distinct batches of 50 examples each, covering… See the full description on the dataset page: https://huggingface.co/datasets/SAWithanage/en-si-parallel-3k.

sourceHugging Facecc-by-4.0updated 4mo agoView on Hugging Face
1likes21downloads
Dataset Card

Dataset Card for en-si-parallel-3k

Dataset Summary

The en-si-parallel-3k dataset is a high-quality, synthetically generated parallel corpus containing 3,000 English-Sinhala translation pairs. It is specifically designed for fine-tuning Large Language Models (LLMs) to enhance English-to-Sinhala translation capabilities and cross-lingual understanding.

Dataset Composition

The dataset is structured into 60 distinct batches of 50 examples each, covering a wide range of domains to ensure model robustness:

  • —News & Information: Climate change, space exploration, and international sports.
  • —Technical & Procedural: Software installation, kitchen recipes, and electronic troubleshooting.
  • —Business & Professional: Corporate communication, auditing, and product launches.
  • —Descriptive & Creative: Narrative descriptions, book summaries, and art critiques.
  • —Everyday Transactional: Travel arrangements, banking, and public services.

Creation Process

This dataset was generated using Gemini 3.5 Flash. To ensure linguistic consistency and high-quality translations, the generation process followed a strict 3-tier hierarchy for terminology:

  1. 1.Tier 1 (Pure Sinhala): Common, natural terms are translated directly into Sinhala.
  2. 2.Tier 2 (Bracket Bridge): Academic or rare terms include the English original in brackets for instant clarity.
  3. 3.Tier 3 (Transliteration/Acronym): Global standards and technical acronyms (e.g., RAM, Wi-Fi) are maintained in Latin characters.

Usage

The dataset is provided in CSV format. You can load it using the Hugging Face datasets library:

python
from datasets import load_dataset

# Load the dataset
dataset = load_dataset("your-username/en-si-parallel-3k")

# Access the data
print(dataset['train'][0])

Licensing and Usage

This dataset is released under the CC-BY-4.0 license, allowing for open use in research and fine-tuning projects.

Limitations and Biases

As this dataset was generated by an AI model, it may contain artifacts or hallucinations typical of synthetic data. While curated for quality, it should be validated for use in mission-critical applications where high-stakes translations are required.