Downloads · 30 days
6
27% of all-time downloads
cohere-transcribe-darija-decoder-lora
01Yassine/cohere-transcribe-darija-decoder-lora
cohere-transcribe-darija-decoder-lora is a automatic speech recognition model from 01Yassine. Use it when you need speech turned into text. It is set up for peft. The card lists the license as other.
Public adapter on Cohere Transcribe Arabic. LoRA on the decoder only. Encoder frozen. Same 3h YouTube set (01Yassine/darija-asr-3h).
All-time downloads
22
Public
Repo size
16.8 MB
Likes
0
Public
Repo makeup
Click a slice to open those files.
.safetensors16.8 MB · 90%
At a glance
- Task
- Automatic Speech Recognition
- Library
- peft
- License
- other
- Access
- Public
- Created
- Aug 30, 2026
- Updated
- Aug 30, 2026
- SHA
- 823173de
Base models
Model card
From the Hugging Face model README
Cohere Transcribe — Moroccan Darija (decoder LoRA)
Public adapter on Cohere Transcribe Arabic. LoRA on the decoder only. Encoder frozen. Same 3h YouTube set (01Yassine/darija-asr-3h).
Eval: atlasia/darija-asr-benchmark (114 clips, human).
| CER | WER | |
|---|---|---|
| base | 20.2 | 49.1 |
| this adapter | 20.2 | 45.1 |
The recommended student is hybrid at 14.4 CER.
Inference from the Hub
from huggingface_hub import snapshot_download
import sys
sys.path.insert(0, snapshot_download("01Yassine/cohere-transcribe-darija-decoder-lora"))
from infer import transcribe
print(transcribe("clip.wav", model_id="01Yassine/cohere-transcribe-darija-decoder-lora"))
| recipe | Hub | AtlasIA CER |
|---|---|---|
| hybrid | 01Yassine/cohere-transcribe-darija | 14.4 |
| full LoRA | 01Yassine/cohere-transcribe-darija-full-lora | 16.5 |
| encoder LoRA | 01Yassine/cohere-transcribe-darija-encoder-lora | 17.4 |
| decoder LoRA | 01Yassine/cohere-transcribe-darija-decoder-lora | 20.2 |
- Task
- Automatic Speech Recognition
- Library
- peft
- License
- other
- Languages
- ar, ary
- Created
- Aug 30, 2026
- Updated
- Aug 30, 2026