---
library_name: transformers.js
base_model: syvai/hviske-v2
tags:
  - onnx
  - whisper
  - transformers.js
  - automatic-speech-recognition
  - speech-to-text
  - danish
  - da
  - webgpu
  - browser
language:
  - da
pipeline_tag: automatic-speech-recognition
license: openrail
---

# Hviske v2 ONNX

ONNX export of [syvai/hviske-v2](https://huggingface.co/syvai/hviske-v2) for use with [Transformers.js](https://huggingface.co/docs/transformers.js) in the browser.

Hviske v2 is a state-of-the-art Danish speech-to-text model, fine-tuned from [openai/whisper-large-v3](https://huggingface.co/openai/whisper-large-v3) on CoRal and Common Voice 17.0 datasets. Developed by [syv.ai](https://syv.ai).

## Performance

| Model | Parameters | CoRal CER | CoRal WER |
|-------|-----------|-----------|-----------|
| **syvai/hviske-v2** | 1540M | **4.7%** | **11.8%** |
| alexandrainst/roest-315 | 315M | 6.6% | 17.0% |
| openai/whisper-large-v3 | 1540M | 11.4% | 28.3% |

## Usage with Transformers.js

```js
import { pipeline } from '@huggingface/transformers'

const transcriber = await pipeline(
  'automatic-speech-recognition',
  'varsan-g/hviske-v2-onnx',
)

const result = await transcriber(audioData)
console.log(result.text)
```

## Model Details

- **Base model:** openai/whisper-large-v3
- **Fine-tuned by:** [syv.ai](https://syv.ai)
- **Language:** Danish (da)
- **License:** OpenRAIL
- **Format:** ONNX (int8 quantized, single-file, no external data)
- **Encoder:** 615 MB
- **Decoder:** 932 MB

## Conversion

Exported from PyTorch using [Optimum](https://huggingface.co/docs/optimum-onnx) and quantized with ONNX Runtime dynamic int8 quantization for browser compatibility:

```bash
optimum-cli export onnx --model syvai/hviske-v2 --task automatic-speech-recognition output/
```

## File Structure

```
onnx/
  encoder_model.onnx           (615 MB)
  decoder_model_merged.onnx    (932 MB)
```