Skip to content

Latest commit

 

History

History
90 lines (54 loc) · 3.13 KB

README.md

File metadata and controls

90 lines (54 loc) · 3.13 KB

faster-whisper-ollama

A real-time voice chatbot based on foster whisper and ollama

Local Audio Recording and Transcription

This script provides a local audio recording and transcription service without relying on cloud services, making it suitable for offline use.

Dependencies

Before running the script, you need to install the following dependencies:

pip3 install pyaudio webrtcvad faster-whisper

Running

To run the script, use the following command:

python3 faster_whisper_demo.py

Overview

This script records audio from the microphone, transcribes it using a local model, and provides real-time transcription services. Below is an overview of the main components:

1. Queues Class

  • Manages two queues: audio for storing audio data and text for storing transcribed text.

2. Transcriber Class

  • Utilizes the FasterWhisper model for audio transcription.
  • Initializes the model with specified parameters such as model size, device, and compute type.
  • Processes audio data and yields transcribed text segments.

3. AudioRecorder Class

  • Records audio using the PyAudio library and processes it with WebRTC VAD (Voice Activity Detection).
  • Initializes the audio stream and VAD, and manages the recording and buffering of audio frames.
  • Detects speech activity and manages the transition between recording and silence.

4. Chat Class

  • Handles communication with the Ollama chatbot.
  • Sends transcribed text to the chatbot and receives responses.

5. main Function

  • Initializes and starts the audio recorder, transcriber, and chat components.
  • Manages the lifecycle of these components and handles exceptions.

Usage

  1. Audio Recording:

    • The script continuously records audio from the microphone.
    • When speech is detected, it starts recording, and when silence is detected, it stops and processes the audio.

  2. Audio Transcription:

    • The recorded audio is sent to the Transcriber class, which uses the FasterWhisper model to transcribe the audio into text.

  3. Chat Interaction:

    • The transcribed text is sent to the Chat class, which communicates with the Ollama chatbot and retrieves responses.

Logging

  • The script uses the logging module to log various stages of the recording, transcription, and chat processes.
  • Log messages provide information on the status and any errors encountered during the execution.

Example Output

  • Transcribed text is logged in real-time, and responses from the chatbot are printed to the console.

Error Handling

  • The script includes error handling to manage exceptions during initialization, audio recording, transcription, and chat communication.

Environment Variable

  • Sets the environment variable KMP_DUPLICATE_LIB_OK to "TRUE" to resolve potential issues with library conflicts.

Notes

  • Ensure that your microphone is properly configured and accessible by the script.
  • The script is designed to work with specific models and configurations. Adjust the parameters as needed for your environment.

License

This project is licensed under the MIT License - see the LICENSE file for details.