Dataset · tarteel-ai
Quranic ayah audio recordings with diacritics for ASR
tarteel-ai/everyayah
Arabic audio corpus of Quranic recitations with precise text transcripts for speech recognition.
What's inside
Dataset from Tarteel AI containing 100k to 1M audio-text pairs in Arabic. Features Quranic ayahs performed by various reciters, with each text transcript provided with full diacritical marks.
What you can do with it
- Training and fine-tuning ASR models for Classical Arabic recognition
- Building automated systems for checking recitation accuracy and diacritics
- Tasks
- Automatic speech recognition
- Languages
- Arabic
- Size
- 100K–1M rows
- Downloads this month
- 4.5K
- Likes
- 46
Models trained on it
- naazimsnh02/whisper-large-v3-turbo-ar-quran 3.1K
- mohammed/fastconformer-quran-arAutomatic speech recognition 2.9K
- nvidia/stt_ar_fastconformer_hybrid_large_pcd_v1.0Automatic speech recognition 2.9K
- baristiran/whisper-turbo-quranAutomatic speech recognition 653
- Hisnlabs/tajweed-v2Automatic speech recognition 393
- deepdml/whisper-base-ar-quran-mix-normAutomatic speech recognition 346
- deepdml/whisper-small-ar-quran-mix-normAutomatic speech recognition 327
- IJyad/whisper-large-v3-TarteelAutomatic speech recognition 191
How to load it
from datasets import load_dataset
ds = load_dataset("tarteel-ai/everyayah")