Configuration Parsing Warning:In adapter_config.json: "peft.task_type" must be a string
Cohere Transcribe — Moroccan Darija (encoder LoRA)
Public adapter on Cohere Transcribe Arabic. LoRA on the Conformer only. Decoder 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 | 17.4 | 47.7 |
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-encoder-lora"))
from infer import transcribe
print(transcribe("clip.wav", model_id="01Yassine/cohere-transcribe-darija-encoder-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 |
- Downloads last month
- 11
Model tree for 01Yassine/cohere-transcribe-darija-encoder-lora
Base model
CohereLabs/cohere-transcribe-03-2026