---
license: apache-2.0
language:
- de
library_name: mlx-audio
base_model: mlx-community/Qwen3-ASR-0.6B-5bit
tags:
- mlx
- mlx-audio
- speech-to-text
- asr
- stt
- medical
- lora
---

# anamedi Personalized ASR

Personalized German medical ASR model, fine-tuned via [anamedi](https://www.anamedi.com) on the doctor's own voice.

**Base model:** [mlx-community/Qwen3-ASR-0.6B-5bit](https://huggingface.co/mlx-community/Qwen3-ASR-0.6B-5bit)  
**Language:** German (de)  
**Domain:** Medical dictation  
**Fine-tuning:** LoRA adapters merged into base weights

## Loading

```python
from mlx_audio.stt.utils import load_model
model = load_model("jempf/554a0b4dc84e4b06")
```

## License

Apache 2.0 — inherited from the base model [Qwen3-ASR](https://huggingface.co/mlx-community/Qwen3-ASR-0.6B-5bit).