PaccyC/kinyarwanda-voice-assistant

Here’s a well-structured `README.md` for your Kinyarwanda Voice Assistant project, using appropriate markdown tags for clarity and documentation best practices: --- ```markdown # 🗣️ Kinyarwanda Voice Assistant A bilingual voice assistant built with Whisper for speech-to-text and TTS for text-to-speech, supporting Kinyarwanda Q&A and voice synthesis. It features an interactive Gradio interface with two main modes: voice assistant and custom text-to-speech (TTS). --- ## 🔧 Features - 🎙️ **Speech Recognition** using OpenAI Whisper. - 💬 **Kinyarwanda Q&A** using predefined text responses. - 🔊 **Text-to-Speech** using Hugging Face's `TTS` API. - 🖥️ **Interactive Interface** using Gradio Tabs. - 🔁 **Speaker Voice Transfer** (using reference audio). - 🌐 **Web UI Launch with Sharing Option**. --- --- ## 🛠️ Installation & Setup ### ⚙️ Python Environment Install dependencies in your Python environment or Jupyter notebook (e.g., Colab): ```bash pip install -q openai-whisper pip install numpy==1.24.3 --force-reinstall pip install gradio pip install transformers pip install torchaudio pip install TTS pip install nemo-toolkit ``` > You can also install all required packages with: ```bash pip install --no-cache-dir -r /content/drive/MyDrive/kinya-assistant/stt/requirements.txt ``` ### 🧰 System Dependencies (for SoX audio tools) ```bash apt-get update && apt-get install -y sox libsox-fmt-all ``` --- ## 🚀 How to Use ### 🎤 Voice Assistant - Upload your audio file or record directly. - The system: - Transcribes audio using Whisper. - Matches text to pre-defined Kinyarwanda questions. - Responds using TTS and plays the audio back. ### 📢 Text-to-Speech - Enter your desired text. - Provide a WAV file as a reference speaker. - Choose a language (`` supported). - Synthesized speech will be generated and played back. --- ## 🧠 Code Overview ### Load Whisper & TTS Models ```python whisper_model = whisper.load_model("small") hf_model = TTS(model_name="tts_models/en/ljs …