This project provides a local, client-side interface for the Orpheus 3B Text-to-Speech model, enabling high-quality speech generation without cloud dependencies. It targets users who want to integrate advanced TTS capabilities into their applications or workflows without relying on external APIs, offering multiple voice options and emotion tags.

How It Works

The project leverages LM Studio to run the Orpheus 3B model locally. It acts as a client that interacts with the LM Studio API, sending text prompts and receiving generated audio. This approach allows for efficient, on-device processing of TTS tasks, benefiting from the model's quality while maintaining user privacy and control.

Quick Start & Requirements

• Install LM Studio and download the orpheus-3b-0.1-ft-q4_k_m.gguf model.
• Load the Orpheus model in LM Studio and start the local server (default: http://127.0.0.1:1234).
• Install dependencies:

  python3 -m venv venv
  source venv/bin/activate
  pip install -r requirements.txt
  

• Run the script: python gguf_orpheus.py --text "Hello, this is a test" --voice tara
• Requires Python 3.x and LM Studio.

Highlighted Details

• Supports multiple distinct voices (tara, leah, jess, leo, dan, mia, zac, zoe).
• Allows audio saving to WAV files.
• Includes support for emotion tags (e.g., <giggle>, <laugh>) within the input text.
• Offers adjustable generation parameters like temperature and repetition penalty.

Maintenance & Community

No specific community channels or maintenance details are provided in the README.

Licensing & Compatibility

Licensed under Apache 2.0, permitting commercial use and modification.

Limitations & Caveats

The project's functionality is entirely dependent on a correctly configured and running LM Studio instance with the Orpheus 3B model loaded. No specific performance benchmarks or system requirements beyond LM Studio are detailed.