🎙️ NeuralChatter TTS 🌍

Multilingual TTS Server with LoRA Fine-Tuning

🔀 Enhanced Fork by Hussein

📺 Demo Video

---

NeuralChatter is an enhanced Chatterbox TTS server with multilingual support, LoRA fine-tuning, and a modern React UI. Generate high-quality speech in 24+ languages with voice cloning and custom voice training capabilities.

---

⭐ What's New in NeuralChatter

Feature	Description
🧠 LoRA Fine-Tuning	Train custom voice models using your own audio data
🌍 Cross-Language Cloning	Train with Arabic audio → Generate speech in English, French, any language!
⚛️ Modern React UI	Beautiful, responsive web interface with Tailwind CSS
📊 Training Scripts	train_lora.py and train_lora_pro.py for voice training
🎙️ Arabic Voice Training	Pre-configured training data for Arabic voice cloning
🔧 Enhanced Configuration	Extended config.py with more customization options

🎯 Clone your voice with just ~10 audio files! Record yourself speaking in Arabic, train the LoRA model, and your cloned voice can speak in 24+ languages!

---

✨ Key Features

From Original Project

• 🌐 24+ Languages: Arabic, Chinese, Danish, Dutch, English, Finnish, French, German, Greek, Hebrew, Hindi, Italian, Japanese, Korean, Malay, Norwegian, Polish, Portuguese, Russian, Spanish, Swedish, Swahili, Turkish
• 🎤 Voice Cloning: Clone any voice using reference audio files
• 📚 Audiobook Generation: Process entire books with automatic text chunking
• 🚀 Easy Windows Installation: One-click setup with install.bat
• ⚡ GPU Acceleration: NVIDIA (CUDA) and AMD (ROCm) support
• 📡 FastAPI Server: RESTful API with interactive documentation

New in NeuralChatter

• 🧠 LoRA Fine-Tuning: Train personalized voice models
• ⚛️ React Web UI: Modern interface built with React + Vite + Tailwind CSS
• 📁 Training Data Management: Organized loRa/ folder for training datasets
• 🔊 Audio Processing: Tools for audio trimming and preprocessing

---

🧠 LoRA Fine-Tuning (New!)

Train custom voice models using your own audio recordings:

Quick Start

# Activate environment
source venv/bin/activate  # Linux
# or
.\venv\Scripts\Activate.ps1  # Windows

# Run training
python train_lora.py --audio_file "path/to/audio.wav" --transcript "path/to/text.txt"

Training Data Structure

Place your training data in the loRa/ folder:

loRa/
├── your_audio.wav          # Audio file (WAV/MP3)
├── your_audio.txt          # Transcript text file
└── your_audio_training_data.json  # Generated training data

---

🚀 Quick Installation (Windows)

1. Download

   git clone https://github.com/husseinnsourr/NeuralChatter.git
   cd NeuralChatter

2. Install

   • Double-click install.bat
   • Choose your hardware:
     • Option 1: NVIDIA GPU (CUDA)
     • Option 2: CPU Only

3. Run

   • Double-click win-run.bat

---

🔧 Configuration

Edit config.yaml for settings:

server:
  host: "0.0.0.0"
  port: 8004

tts_engine:
  device: "auto" # auto, cuda, cpu

---

📜 License

This project is licensed under the GNU General Public License v3.0 (GPL-3.0). Contributions must be shared back with the community.

---

🤝 Contributing

Contributions welcome! Report bugs, suggest features, or submit pull requests. Open issue here.

---

🎙️ Generate speech in 24+ languages with LoRA fine-tuning and one-click Windows installation!