Dataset Source:

Original Source: The English sentences were sourced from oubic domain sources such as https://www.gutenberg.org/ .
Translation Tool: Google Translate was used for translating the sentences from English to Yoruba.

Dataset Format:

english: The original English sentence.
yoruba: The Yoruba translation of the sentence.
source: the source of the English sentences.

Example:

en	yo	source
The subconscious offensiveness of their attitude has constituted old Jolyon's 'home' the psychological moment of the family history, made it the prelude of their drama.	Iwa ibinu Ã¨roÅ„gbÃ ti iá¹£esi wá»n ti jáº¹ â€œileâ€ atijá» ti Jolyon ni akoko imá»-jinláº¹ ti itan-aká»á»láº¹ áº¹bi, jáº¹ ki o jáº¹ iá¹£aaju ti erÃ© wá»n.	https://www.gutenberg.org/ebooks/2559.txt.utf-8
The Forsytes were resentful of something, not individually, but as a family; this resentment expressed itself in an added perfection of raiment, an exuberance of family cordiality, an exaggeration of family importance, and--the sniff.	Awá»n Forsytes binu si nkan kan, kii á¹£e olukuluku, á¹£ugbá»n gáº¹gáº¹bi idile; ibinu yii á¹£e afihan araráº¹ ni pipe ti aá¹£á» ti a fi kun, igbadun ti ifarabaláº¹ idile, iá¹£aju ti pataki idile, ati --ifun.	https://www.gutenberg.org/ebooks/2559.txt.utf-8
Danger--so indispensable in bringing out the fundamental quality of any society, group, or individual--was what the Forsytes scented; the premonition of danger put a burnish on their armour.	Ewu - nitorinaa ko á¹£e pataki lati mu didara ipiláº¹ ti awujá», áº¹gbáº¹, tabi áº¹ni ká»á»kan jade - jáº¹ ohun ti awá»n Forsytes rÃ¹n; premonition ti ewu fi kan inÃ¡ lori wá»n ihamá»ra.	https://www.gutenberg.org/ebooks/2559.txt.utf-8

Dataset Size:

Number of Entries: 520,000

Usage:

This dataset can be used for:

Training machine translation models for Yoruba.
Analyzing translation quality and limitations in automated tools.
Supporting linguistic research and NLP projects for low-resource languages.

Limitations and Considerations:

Quality of Translations: As translations were generated using Google Translate, some sentences may not reflect perfect accuracy. Manual validation is recommended for critical applications.
Cultural and Contextual Nuances: Machine translations might miss idiomatic expressions or cultural nuances present in the source language.
Biases: Any biases inherent in Google Translate's model may propagate into this dataset.

Licensing:

Source Material License: Public Domain

Tags:

machine-translation
speech-to-text
yoruba-language
african-languages

Task_categories:

text-classification
machine-translation