Arabic Audio Dataset with Text Modality

Over 100,000 audio clips, likely up to 1 million, form this collection focused on Arabic speech and audio. The dataset was created by MohamedGomaa30 and was last updated in February 2026. It includes a text modality, suggesting paired audio and transcript data.

Use Cases

• Train an automatic speech recognition model to convert Arabic audio clips into text transcripts.
• Develop an audio classification model to identify speaker characteristics or acoustic events using audio features.
• Fine-tune a multimodal model for tasks requiring joint understanding of Arabic audio and its corresponding text data.
• Benchmark model performance on Arabic-language audio processing tasks using a standardized dataset.

Strengths

• Dataset size is categorized as between 100K and 1M samples, indicating a substantial collection.
• Includes both audio and text modalities, enabling multimodal machine learning applications.
• Data is stored in an optimized Parquet format, which is efficient for large-scale processing.

Limitations

• Specific column names, audio duration, sample rate, and transcription quality are unknown.
• The geographic focus appears to be on the 'us' region, which may limit dialectal diversity for Arabic speech.
• License terms are unspecified, creating uncertainty for commercial use and redistribution.

Provenance

Source

Hugging Face, uploaded by author MohamedGomaa30.

Collection Method

Collection method is unspecified but likely involves web scraping or aggregation from existing sources.

Freshness

Last updated in February 2026, indicating recent maintenance.

Geography

Region tag suggests a primary focus on the United States ('us').