English-Korean Aligned Documentation for Transformers Library

Aligned text pairs from the English and Korean documentation of the Hugging Face Transformers library. The dataset was created by author jmj-minju and was last updated in April 2026. It contains a training split for machine translation tasks.

Use Cases

• Train a neural machine translation model to translate English technical documentation to Korean using the aligned text pairs.
• Fine-tune a multilingual language model on the parallel corpus to improve its performance on technical NLP terminology.
• Analyze the alignment quality and translation patterns between English source files and their Korean counterparts.

Strengths

• Data is derived from the official Hugging Face Transformers documentation, ensuring authoritative source material.
• Provides a structured training split (train.jsonl) specifically prepared for model development.

Limitations

• The total number of rows, sentences, or word count in the dataset is unknown.
• The dataset's scope is limited to the specific domain of machine learning library documentation, which may not generalize to other text types.

Provenance

Source

Documentation files from the huggingface/transformers GitHub repository (docs/source/en and docs/source/ko trees).

Collection Method

Text pairs were algorithmically aligned from the source documentation trees.

Freshness

Last updated on 2026-04-05.