Downloads · 30 days
220
31% of all-time downloads
rabah2026/wav2vec2-large-xlsr-53-arabic-quran
wav2vec2-large-xlsr-53-arabic-quran is a automatic speech recognition model from rabah2026. Use it when you need speech turned into text. The card lists the license as apache-2.0.
Ce modèle est une version fine-tunée de jonatasgrosman/wav2vec2-large-xlsr-53-arabic sur le dataset Quran Ayah Corpus (Rabah2026).
Downloads · 30 days
220
31% of all-time downloads
All-time downloads
700
Public
Parameters
315M
1.3 GB on disk
Likes
0
Public
Click a slice to open those files.
.safetensors1.3 GB · 100%
From the Hugging Face model README
Ce modèle est une version fine-tunée de jonatasgrosman/wav2vec2-large-xlsr-53-arabic sur le dataset Quran Ayah Corpus (Rabah2026).
Il est spécialisé pour la reconnaissance vocale du Coran (Tajweed & Murattal) et supporte l'écriture Uthmani (incluant les caractères spéciaux comme ٱ et ٰ).
from transformers import Wav2Vec2ForCTC, Wav2Vec2Processor
import torch
import librosa
model_id = "VOTRE_NOM_UTILISATEUR/wav2vec2-large-xlsr-53-arabic-quran"
processor = Wav2Vec2Processor.from_pretrained(model_id)
model = Wav2Vec2ForCTC.from_pretrained(model_id)
# Chargement audio (16kHz obligatoire)
audio, sr = librosa.load("verset.mp3", sr=16000)
inputs = processor(audio, sampling_rate=16000, return_tensors="pt", padding=True)
with torch.no_grad():
logits = model(inputs.input_values).logits
predicted_ids = torch.argmax(logits, dim=-1)
transcription = processor.batch_decode(predicted_ids, skip_special_tokens=True)[0]
print(transcription)