Moroccan Darija Speech Dataset with 12,743 Parallel Text and Audio Samples

12,743 parallel text and speech samples for Moroccan Darija, including transcriptions in Latin and Arabic scripts and English translations. It was created by atlasia to support speech recognition, language modeling, and NLP tasks for Moroccan Darija. The dataset was last updated on February 7, 2025.

Use Cases

• Train automatic speech recognition models based on the parallel audio and text samples.
• Develop language models for Moroccan Darija based on the provided transcriptions.
• Build machine translation systems between Moroccan Darija and English using the provided translations.
• Study code-switching or script variation using the Latin and Arabic script transcriptions.

Strengths

• Contains 12,743 parallel samples, providing a foundation for model training.
• Offers transcriptions in both Latin and Arabic scripts, capturing different writing conventions.
• Includes English translations, enabling cross-lingual tasks.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is known, but other scale details like audio duration or file formats are unknown.
• Description metadata is limited; actual data quality requires manual inspection after download.

Provenance

Source

Originally sourced from a repository linked in the description.

Freshness

Last updated 2025-02-07 14:36:46; freshness should be verified.

Geography

Morocco