English-Somali-Parallel-Corpus

This dataset contains high-quality parallel sentence pairs, multi-sentence alignments, and paragraph-level translations between Somali and English. It is specifically curated and structured to train, fine-tune, and evaluate machine translation models and large language models (LLMs) for Somali-English language tasks.

Dataset Details

• Language Pair: Somali (so) <-> English (en)
• Format: JSON Lines (.jsonl)
• Data Structure: Each record contains a JSON object with aligned English and Somali text fields.

Data Fields

• en: The English translation text string.
• so: The corresponding Somali translation text string.

Intended Uses

• Machine Translation (MT) fine-tuning.
• Cross-lingual representation learning.
• Evaluation of LLMs on Somali translation tasks.