Downloads · 30 days
18
0% of all-time downloads
hubert-darija-combined
amineouaki/hubert-darija-combined
hubert-darija-combined is a automatic speech recognition model from amineouaki. Use it when you need speech turned into text. The card lists the license as apache-2.0.
Fine-tuning de omarxadel/hubert-large-arabic-egyptian pour la reconnaissance automatique de la parole (ASR) en darija (dialecte marocain), sur un ensemble combiné de datasets publics.
All-time downloads
6.2K
Public
Parameters
316M
1.3 GB on disk
Likes
1
Public
Repo makeup
Click a slice to open those files.
.safetensors1.3 GB · 100%
At a glance
- Task
- Automatic Speech Recognition
- License
- apache-2.0
- Model type
- hubert
- Access
- Public
- Created
- Jul 26, 2026
- Updated
- Aug 2, 2026
- SHA
- dd679ce0
Base models
Datasets
Model card
From the Hugging Face model README
hubert-darija-combined
Fine-tuning de omarxadel/hubert-large-arabic-egyptian pour la
reconnaissance automatique de la parole (ASR) en darija (dialecte marocain),
sur un ensemble combiné de datasets publics.
Résultats finaux
| Métrique | Valeur |
|---|---|
| Eval WER (Word Error Rate) | 0.2386 (23.86%) |
| Eval CER (Character Error Rate) | 0.0813 (8.13%) |
| Train loss (moyenne sur tout l'entraînement) | 0.41482063716298273 |
Hyperparamètres d'entraînement
| Hyperparamètre | Valeur |
|---|---|
| Modèle de base | omarxadel/hubert-large-arabic-egyptian |
| Époques | 50.0 |
| Batch size (par device) | 2 |
| Gradient accumulation | 16 |
| Batch effectif | 32 |
| Learning rate | 5e-5 |
| Weight decay | 0.005 |
| Optimiseur | AdamW (par défaut Trainer) |
| Précision | fp16 |
| Gradient checkpointing | Oui (use_reentrant=False) |
| Feature encoder | gelé (freeze_feature_encoder) |
| Attention dropout | 0.1 |
| Hidden dropout | 0.1 |
| Feat proj dropout | 0.0 |
| Mask time prob | 0.05 |
| LayerDrop | 0.1 |
| CTC loss reduction | mean |
| Durée d'entraînement | 1 day(s), 01:51:47 |
| Vitesse | 21.547 samples/s, 0.673 steps/s |
Datasets utilisés
adiren7/darija_speech_to_textmohamedmou/moroccan-darija-asr-dataset-splitafyfbadreddine77/darija-asr-datasetntariklk/darija-merged-asratlasia/Moroccan-Darija-Wiki-Audio-DatasetDatasmartly/moroccan_darija_audio (accès refusé — dataset privé, non inclus dans l'entraînement)RHEZLOUNE/darija-restaurant-audioanaszil/Segmented-Moroccan-Darija-Wiki-Audio-Dataset
Filtre appliqué : durée audio ≤ 15s.
Utilisation
from transformers import HubertForCTC, Wav2Vec2Processor
import librosa
import torch
processor = Wav2Vec2Processor.from_pretrained("amineouaki/hubert-darija-combined")
model = HubertForCTC.from_pretrained("amineouaki/hubert-darija-combined")
speech, _ = librosa.load("audio.wav", sr=16000)
inputs = processor(speech, sampling_rate=16000, return_tensors="pt", padding=True)
with torch.no_grad():
logits = model(**inputs).logits
pred_ids = torch.argmax(logits, dim=-1)
print(processor.decode(pred_ids[0]))
Auteur
Amine Ouakib
- Email : [email protected]
- GitHub : ouakibamine
- LinkedIn : amine-ouakib
- Task
- Automatic Speech Recognition
- Type
- hubert
- License
- apache-2.0
- Languages
- ar
- Created
- Jul 26, 2026
- Updated
- Aug 2, 2026