Multi-Language Support (Italian/Spanish)

This repository also includes example scripts for running fine-tuned models, such as lex-au/Orpheus-3b-Italian_Spanish-FT-Q8_0.gguf.

To use a different language model:

1. Stop the LM Studio server and load your desired model.
2. Start the server again.
3. Use the dedicated scripts (italian_orpheus.py, italian_structured_audiobook_generator.py) which are pre-configured with the correct voice prompts.
4. EXAMPLE FOR ITALIAN: python italian_orpheus.py --text "Ciao mondo, questo è un test." --voice it-male

Orpheus-TTS-Local

A lightweight client for running Orpheus TTS locally using LM Studio API.

Features

• 🎧 High-quality Text-to-Speech using the Orpheus TTS model
• 💻 Completely local - no cloud API keys needed
• 🔊 Multiple voice options (tara, leah, jess, leo, dan, mia, zac, zoe)
• 💾 Save audio to WAV files
• 📚 Generate audiobooks from text, markdown, and HTML files
• 🧠 Intelligent text chunking using spaCy for natural-sounding speech

Quick Setup

1. Install LM Studio

2. Download the Orpheus TTS model (orpheus-3b-0.1-ft-q4_k_m.gguf) in LM Studio

3. Load the Orpheus model in LM Studio

4. Start the local server in LM Studio (default: http://127.0.0.1:1234)

5. Install dependencies:

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

6. Run the script:

   python gguf_orpheus.py --text "Hello, this is a test" --voice tara
   

   Note: When running the audiobook generators for the first time, they will automatically download the required spaCy language model if it's not already installed.

Usage

Basic Text-to-Speech

python gguf_orpheus.py --text "Your text here" --voice tara --output "output.wav"

Options

• --text: The text to convert to speech
• --voice: The voice to use (default: tara)
• --output: Output WAV file path (default: auto-generated filename)
• --list-voices: Show available voices
• --temperature: Temperature for generation (default: 0.6)
• --top_p: Top-p sampling parameter (default: 0.9)
• --repetition_penalty: Repetition penalty (default: 1.1)

Audiobook Generation

Simple Audiobook Generator

python audiobook_generator.py mybook.txt --voice leo --output-dir "my_audiobooks"

Options

• input_file: Input text, markdown, or HTML file
• --voice: Voice to use (default: tara)
• --output-dir: Output directory for audio files (default: outputs/audiobook)
• --separate-chunks: Keep audio chunks as separate files
• --by-structure: Chunk by document structure (for markdown/html)
• --list-voices: List available voices and exit

Structured Audiobook Generator

For more advanced document processing with better structure handling:

python structured_audiobook_generator.py mybook.md --voice tara --max-length 600

Options

• input_file: Input text, markdown, or HTML file
• --voice: Voice to use (default: tara)
• --output-dir: Output directory for audio files (default: outputs/audiobook)
• --separate-chunks: Keep audio chunks as separate files
• --max-length: Maximum characters per chunk (default: 500)
• --min-length: Minimum characters per chunk (default: 50)
• --list-voices: List available voices and exit

Using as a Library

See example.py for basic TTS usage and example_audiobook.py for audiobook generation examples.

Available Voices

• tara - Best overall voice for general use (default)
• leah
• jess
• leo
• dan
• mia
• zac
• zoe

Emotion

You can add emotion to the speech by adding the following tags:

<giggle>
<laugh>
<chuckle>
<sigh>
<cough>
<sniffle>
<groan>
<yawn>
<gasp>

License

Apache 2.0