EGYSpeak: 147,979 Egyptian Arabic Speech Clips with Transcriptions

EGYSpeak is a curated dataset of 147,979 single-speaker Egyptian Arabic audio clips paired with transcriptions. It was created by MohamedGomaa30, sourced from the fadisarwat/egyptian-arabic-lines Kaggle dataset and processed through an ASR pipeline. The dataset was last updated on Hugging Face in April 2026.

Use Cases

• Train automatic speech recognition models based on the Egyptian Arabic audio and transcription pairs.
• Fine-tune speech synthesis models using the single-speaker, dialect-specific audio clips.
• Benchmark dialectal Arabic NLP systems on the provided transcriptions.
• Conduct linguistic analysis of Egyptian Arabic features mentioned in the description.

Strengths

• Contains 147,979 audio samples, providing a substantial corpus for model training.
• Focuses on a single-speaker and pure Egyptian Arabic dialect, which may enhance consistency.
• Data was processed through a rigorous ASR pipeline, suggesting quality control.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is known, but other metadata like file formats, size, and license is unknown.
• Data may reflect source bias inherent to the original Kaggle dataset and processing pipeline.

Provenance

Source

Sourced from the fadisarwat/egyptian-arabic-lines Kaggle dataset.

Collection Method

Processed through a rigorous ASR pipeline.

Time Range

null

Freshness

Last updated 2026-04-28 22:33:55; freshness should be verified.

Geography

Egyptian Arabic dialect suggests a primary geographic focus on Egypt.