File size: 3,575 Bytes
c783336
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
---
license: cc-by-4.0
language:
- en
library_name: gguf
tags:
- automatic-speech-recognition
- medical
- parakeet
- gguf
- parakeet.cpp
- omi-med-stt
pipeline_tag: automatic-speech-recognition
base_model: nvidia/parakeet-tdt-0.6b-v2
---

# Omi Med STT v1 GGUF

GGUF export of [Omi Med STT v1](https://huggingface.co/omi-health/omi-med-stt-v1)
for Linux and Windows CPU use through the `omi-med-stt` CLI.

This is the portability path. If you have Apple Silicon, use the MLX q8 repo. If
you have an NVIDIA GPU, use the canonical NeMo checkpoint.

## Quickstart

```bash
pip install -U omi-med-stt
omi-med-stt install-cpp --cpp-backend cpu
omi-med-stt audio.wav --runtime cpp
```

## Files

| File | Status |
|---|---|
| `omi-med-stt-v1-q8_0.gguf` | Default CPU artifact, benchmarked |
| `omi-med-stt-v1-f16.gguf` | Provided for conversion/experimentation; not independently benchmarked |

## Evaluation

Full evaluation details: [omi.health/research/omi-med-stt](https://omi.health/research/omi-med-stt/).
Benchmark: 7.18h of real and synthetic clinical speech across dialogue, dictation, medication review, procedures/devices/tests, and general speech. Speed is shown as time to process one hour of audio; lower is faster.

### NeMo vs Open / Local Models

Local GPU baselines were run on A10 where applicable; VibeVoice-ASR 9B used H100.

| Model | WER | M-WER | Drug M-WER | Medical Recall | Speed: time / 1 hour audio (formula-derived x realtime) |
|---|---:|---:|---:|---:|---:|
| VibeVoice-ASR 9B | 11.10% | 1.78% | 1.36% | 98.71% | 5m 20s (11.2x) |
| **Omi Med STT v1 NeMo** | **8.30%** | **2.37%** | **4.75%** | **97.95%** | **25s (146.3x)** |
| Qwen3 ASR 1.7B | 10.72% | 3.13% | 6.11% | 97.21% | 44s (81.1x) |
| Whisper Large v3 Turbo (A10) | 11.98% | 3.93% | 5.88% | 96.45% | 1m 19s (45.8x) |
| Cohere Transcribe 03-2026 | 14.88% | 5.05% | 11.09% | 95.16% | 25s (146.3x) |
| Parakeet TDT 0.6B v3 | 15.26% | 8.01% | 9.50% | 96.34% | 23s (157.9x) |
| Parakeet TDT 0.6B v2 base | 16.45% | 8.36% | 8.60% | 96.20% | 23s (153.8x) |

### Runtime Artifacts

Same internal evaluation as the canonical checkpoint.

| Artifact | WER | M-WER | Drug M-WER | Medical Recall | Speed: time / 1 hour audio (formula-derived x realtime) |
|---|---:|---:|---:|---:|---:|
| NeMo canonical | 8.30% | 2.37% | 4.75% | 97.95% | 25s (146.3x) |
| MLX q8 | 8.61% | 2.75% | 5.20% | 97.63% | 53s (67.4x) |
| **GGUF q8_0** | **9.12%** | **3.20%** | **6.33%** | **97.53%** | **2m 53s (20.8x)** |

The GGUF q8_0 build is useful when CPU portability matters. It is not the
quality-leading artifact.

## Compatibility

These files are **not llama.cpp text-model GGUF files**. They require a Parakeet
ASR runtime. The supported path is:

```bash
omi-med-stt audio.wav --runtime cpp
```

The CLI installs the patched `parakeet.cpp` runtime needed for Omi Med STT v1.

## Links

- Canonical model: [`omi-health/omi-med-stt-v1`](https://huggingface.co/omi-health/omi-med-stt-v1)
- Mac q8 default: [`omi-health/omi-med-stt-v1-mlx-q8`](https://huggingface.co/omi-health/omi-med-stt-v1-mlx-q8)
- Runtime CLI: [`Omi-Health/omi-med-stt-runtime`](https://github.com/Omi-Health/omi-med-stt-runtime)
- Broader evaluation and product context: [omi.health/research/omi-med-stt](https://omi.health/research/omi-med-stt/)
- parakeet.cpp: [`mudler/parakeet.cpp`](https://github.com/mudler/parakeet.cpp)

## Safety

Omi Med STT v1 is speech-to-text only. It is not a diagnostic, triage,
prescribing, or clinical decision model, and it is not clinically validated.
Transcripts must be reviewed before any clinical use.