A speech transcription and translation application using whisper AI model.
- Speech to text
- Translation of transcribed text (Speech to translated text)
- Realtime input from mic and speaker
- Batch file processing with timestamp
- Whisper uses vram/gpu to process the audio, so it is recommended to have a CUDA compatible GPU. If there is no compatible GPU, the application will use the CPU to process the audio (This might make it slower). For each model requirement you can check directly at the whisper repository or you can hover over the model selection in the app (there will be a tooltip about the model info).
- Speaker input only work on windows 8 and above.
- Download the latest release here
- Install
- Run the program
- Set user setting
- Select model
- Select mode and language
- Click the record button
- Stop record
- (Optionally) export the result to a file
You can change the settings by clicking the settings button on the menubar of the app. Alternatively, you can press F2 to open the menu window or you could also edit the settings file manually located at ./setting/setting.json
.
Warning
As of right now (4th of November 2022) I guess pytorch is not compatible with python 3.11 so you can't use python 3.11. I tried with 3.11 but it doesn't work so i rollback to python 3.10.8.
Note
It is recommended to create a virtual environment, but it is not required.For OS other than windows, you can install the packages from requirements_notwindows.txt
The master branch might not always be stable so you can checkout to the latest release tag to get the latest stable version.
- Create your virtual environment by running
python -m venv venv
- Activate your virtual environment by running
source venv/bin/activate
- Install all the dependencies needed by running the
devSetup.py
located in root directory or install the packages yourself by installing from the requirements.txt yourself by runningpip install -r requirements.txt
- Get to root directory and Run the script by typing
python Main.py
Whisper needs ffmpeg to work, you can install it and add it to your path manually or you can do it easily by running the following command:
# on Ubuntu or Debian
sudo apt update && sudo apt install ffmpeg
# on Arch Linux
sudo pacman -S ffmpeg
# on MacOS using Homebrew (https://brew.sh/)
brew install ffmpeg
# on Windows using Chocolatey (https://chocolatey.org/)
choco install ffmpeg
# on Windows using Scoop (https://scoop.sh/)
scoop install ffmpeg
Note
This process could be handled automatically by running devSetup.py
To use GPU you first need to uninstall torch
then you can go to pytorch official website to install the correct version of pytorch
with GPU compatibily for your system.
You can use pyinstaller or auto-py-to-exe for a graphical interface.
-
If you use pyinstaller you can load the spec file by running
pyinstaller ./build.spec
to build the project. Alternatively, you can type the build command when inroot directory
directly like this:pyinstaller --noconfirm --onedir --console --icon "./assets/icon.ico" --name "Speech Translate" --clean --add-data "./assets;assets/" --copy-metadata "tqdm" --copy-metadata "regex" --copy-metadata "requests" --copy-metadata "packaging" --copy-metadata "filelock" --copy-metadata "numpy" --copy-metadata "tokenizers" --add-data "./venv/Lib/site-packages/whisper/assets;whisper/assets/" "./Main.py"
This will produce an exceutable file in the
dist
directory.Note: Replace the venv with your actual venv path
-
If you use auto-py-to-exe you can load the build.json file located in root directory. You will need to replace the dot (.) in the build.json file with the actual path of the project. This will produce an exceutable file in the
output
directory.
You should be able to compile it on other platform (mac/linux) but I only tested it on Windows.
This project should be compatible with Windows (preferrably windows 10 or later) and other platforms. But I haven't tested it on platform other than windows.
Feel free to contribute to this project by forking the repository, making your changes, and submitting a pull request. You can also contribute by creating an issue if you find a bug or have a feature request. Also, feel free to give this project a star if you like it.
This project is licensed under the MIT License - see the LICENSE file for details
Check out my other similar project called Screen Translate a screen translator / OCR tools made possible using tesseract.