| Author | HoshinoKun |
|---|---|
| [email protected] |
This is a whisper based on the ctranslate2 fine-tuned version that significantly reduces memory usage and optimizes speed. The code was partially referenced from whisper-ctranslate2 and compiled as a native version using nuitka, and is now released on Releases.
Download from Releases
fast-whisper2ass.exe audio --modle base
git clone https://github.com/hoshinohikari/whisper.ctranslate2.git
cd whisper.ctranslate2
pip install -r requirements.txt
python fast-whisper2ass.py audio --modle base
The Releases version includes the CUDA runtime library. If you need to run the source code version, please install CUDA 11 and cuDNN 8 on your own.
If you encounter any of these problems during use, you can try these solutions
- FileNotFoundError
Please download an ffmpeg.exe and place it in the same directory as fast-whisper2ass.exe.