File size: 1,188 Bytes
e8c7d29 291488c e8c7d29 |
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 |
---
license: apache-2.0
language:
- ja
tags:
- automatic-speech-recognition
---
# reazonspeech-k2-v2
`reazonspeech-k2-v2` is an automatic speech recognition (ASR) model
trained on [ReazonSpeech v2.0 corpus](https://huggingface.co/datasets/reazon-research/reazonspeech).
This model provides end-to-end Japanese speech recognition based on
[Next-gen Kaldi](https://k2-fsa.org/).
## Model Architecture
* Character-based RNN-T model. The total parameter count is 159.34M.
* This model utilizes an enhanced Transformer architecture called
[Zipformer](https://arxiv.org/abs/2310.11230).
* The training recipe is available on
[k2-fsa/icefall](https://github.com/k2-fsa/icefall/tree/master/egs/reazonspeech/ASR).
Note that this model can process Japanese audio clips up to ~30 seconds.
## Usage
We recommend to use this model through our
[reazonspeech](https://github.com/reazon-research/reazonspeech)
library.
```
from reazonspeech.k2.asr import load_model, transcribe, audio_from_path
audio = audio_from_path("speech.wav")
model = load_model()
ret = transcribe(model, audio)
print(ret.text)
```
## License
[Apaceh Licence 2.0](https://choosealicense.com/licenses/apache-2.0/)
|