File size: 2,291 Bytes
76bab86 01a8664 76bab86 01a8664 76bab86 01a8664 76bab86 01a8664 76bab86 01a8664 76bab86 01a8664 76bab86 01a8664 76bab86 01a8664 76bab86 01a8664 76bab86 01a8664 76bab86 01a8664 76bab86 01a8664 76bab86 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 | ---
language:
- ar
- ary
license: other
base_model: CohereLabs/cohere-transcribe-arabic-07-2026
library_name: peft
tags:
- asr
- darija
- moroccan-arabic
- speech-recognition
- lora
pipeline_tag: automatic-speech-recognition
---
# Cohere Transcribe — Moroccan Darija (hybrid)
Public adapter on [Cohere Transcribe Arabic](https://huggingface.co/CohereLabs/cohere-transcribe-arabic-07-2026). **Hybrid:** MultiConv on encoder layers 15–47 + LoRA on the decoder. Trained on 3h YouTube Darija ([`01Yassine/darija-asr-3h`](https://huggingface.co/datasets/01Yassine/darija-asr-3h)).
Eval: [`atlasia/darija-asr-benchmark`](https://huggingface.co/datasets/atlasia/darija-asr-benchmark) (114 clips, human).
| | CER | WER |
| --- | ---: | ---: |
| base | 20.2 | 49.1 |
| **hybrid (this repo)** | **14.4** | **38.3** |
You need the Cohere base weights (and its license). This repo is only the adapter.
## Inference (from the Hub, no training clone)
```bash
pip install "transformers>=5.4" peft torch torchaudio soundfile huggingface_hub
```
```python
from huggingface_hub import snapshot_download
import sys
sys.path.insert(0, snapshot_download("01Yassine/cohere-transcribe-darija"))
from infer import transcribe
print(transcribe("clip.wav"))
```
Or, if you already have `infer.py` from this repo:
```bash
python infer.py clip.wav --model hybrid
python infer.py clip.wav --model 01Yassine/cohere-transcribe-darija
```
## Other open checkpoints
Same data and seed, different trainable slice:
| recipe | Hub | AtlasIA CER |
| --- | --- | ---: |
| **hybrid** (MultiConv + LoRA) | [`01Yassine/cohere-transcribe-darija`](https://huggingface.co/01Yassine/cohere-transcribe-darija) | **14.4** |
| full LoRA | [`01Yassine/cohere-transcribe-darija-full-lora`](https://huggingface.co/01Yassine/cohere-transcribe-darija-full-lora) | 16.5 |
| encoder LoRA | [`01Yassine/cohere-transcribe-darija-encoder-lora`](https://huggingface.co/01Yassine/cohere-transcribe-darija-encoder-lora) | 17.4 |
| decoder LoRA | [`01Yassine/cohere-transcribe-darija-decoder-lora`](https://huggingface.co/01Yassine/cohere-transcribe-darija-decoder-lora) | 20.2 |
```python
print(transcribe("clip.wav", model_id="full_lora"))
print(transcribe("clip.wav", model_id="01Yassine/cohere-transcribe-darija-encoder-lora"))
```
|