Logo Lanfrica

yvesHakizimana/kinyarwanda-voice-assistant

Domain:

natural language processing

Record type:

software
Creator:
yve
Host:
an intermediate basic kinyarwanda voice assistant powered NLP which can hear and answer questions by speaking. # Kinya Voice Assistant 🎤 A sophisticated voice assistant that understands and responds in Kinyarwanda, providing a natural conversational interface through speech-to-text and text-to-speech capabilities. ## 🌟 Features - **Voice Input Processing**: Captures and processes Kinyarwanda speech input - **Speech-to-Text (STT)**: Accurate transcription of Kinyarwanda speech using NeMo-based models - **Natural Language Processing**: Handles queries using both rule-based patterns and ChatGPT integration - **Text-to-Speech (TTS)**: High-quality Kinyarwanda speech synthesis using MB-iSTFT-VITS2 - **Interactive Interface**: User-friendly Gradio web interface for easy interaction ## 📝 Description Kinya Voice Assistant is an innovative project that creates a seamless voice interaction experience in Kinyarwanda. The system follows a three-step process: 1. **Speech Recognition**: Converts spoken Kinyarwanda into text 2. **Query Processing**: Analyzes the text and generates appropriate responses 3. **Speech Synthesis**: Converts the response back into natural-sounding Kinyarwanda speech ## 🖼️ Demo ### Terminal Execution ### User Interface ### Usage Example ## 🛠️ Technical Architecture ### Speech-to-Text (STT) - Built using NVIDIA NeMo framework - Utilizes a pretrained Kinyarwanda model from RW-DEEPSPEECH-API - Handles various audio inputs and converts them to text format ### Natural Language Processing - Implements a hybrid approach: - Rule-based pattern matching for common queries - ChatGPT integration for handling complex queries - Custom intent recognition system - Supports various conversation types (greetings, questions, general dialogue) ### Text-to-Speech (TTS) - Based on KinyaTTS - Uses MB-iSTFT-VITS2 model architecture - Generates natural-sounding Kinyarwanda speech ## 🔧 Setup and Installation ### Prerequisites - Python 3.x - NVIDIA GPU (recommended for optimal performance) - Google Colab (for running the notebook) ### Required Packages ```bash pip insta …