This application is a real-time speech-to-text transcription tool that uses the Faster-Whisper model for transcription and the TranslatePy library for translation. The transcribed and translated content is shown in a semi-transparent pop-up window. The Faster-Whisper model enables efficient speech recognition even on devices with 6GB or less VRAM.
To install the dependencies, simply run:
pip install -r requirements.txt
Run the translation_demo.py script with the desired arguments to start the real-time transcription and translation. Example command and arguments can be found in the file.
Transcription text and translations will appear in a pop-up window as audio input is detected.
The application allows a range of options for customizing the transcription process, including choosing the model, destination language, microphone options, and more. Check the arguments in translation_demo.py and modify them as necessary to meet your needs.
The TranscriptionWindow.py script contains a TranscriptionWindow class, which is responsible for displaying the transcription and translation results in a Tkinter-based pop-up window. The update_text method of this class handles the translation and display of text in the window.
This demo was created by modifying the whisper_real_time repository and uses the Faster-Whisper model and the TranslatePy library.