README.md

---
license: mit
language:
- ja
base_model:
- kotoba-tech/kotoba-whisper-v2.2
pipeline_tag: automatic-speech-recognition
---

# Whisper kotoba-whisper-v2.2 model for CTranslate2
This repository contains the conversion of [kotoba-tech/kotoba-whisper-v2.2](https://huggingface.co/kotoba-tech/kotoba-whisper-v2.2) to the [CTranslate2](https://github.com/OpenNMT/CTranslate2) model format.  
This model can be used in CTranslate2 or projects based on CTranslate2 such as [faster-whisper](https://github.com/systran/faster-whisper).

# Example
Install library and get a sample audio.
```python
pip install faster-whisper
```

Inference with the kotoba-whisper-v2.2-faster.
```python
from faster_whisper import WhisperModel

if __name__ == '__main__':
    model = WhisperModel(model_size_or_path="./kotoba-whisper-v2.2-faster", device="cuda", compute_type="float32", local_files_only=True)
    segments, info = model.transcribe(audio="./123.wav", language="ja", chunk_length=5, condition_on_previous_text=False, hotwords="ノイミー")

    for segment in segments:
        print("[%.2fs -> %.2fs] %s" % (segment.start, segment.end, segment.text))

```

# Conversion details
The original model was converted with the following command:
```python
ct2-transformers-converter --model kotoba-tech/kotoba-whisper-v2.2 --output_dir kotoba-whisper-v2.2-faster --copy_files tokenizer.json preprocessor_config.json --quantization float32
```
Note that the model weights are saved in FP32. This type can be changed when the model is loaded using the [compute_type option in CTranslate2](https://opennmt.net/CTranslate2/quantization.html).

# More information
For more information about the kotoba-whisper-v2.0, refer to the original [model card](https://huggingface.co/kotoba-tech/kotoba-whisper-v2.2).