OuteTTS-0.2-500M

Model Description

OuteTTS-0.2-500M is our improved successor to the v0.1 release. The model maintains the same approach of using audio prompts without architectural changes to the foundation model itself. Built upon the Qwen-2.5-0.5B, this version was trained on larger and more diverse datasets, resulting in significant improvements across all aspects of performance.

Special thanks to Hugging Face for providing GPU grant that supported the training of this model!

Key Improvements

• Enhanced Accuracy: Significantly improved prompt following and output coherence compared to the previous version
• Natural Speech: Produces more natural and fluid speech synthesis
• Expanded Vocabulary: Trained on over 5 billion audio prompt tokens
• Voice Cloning: Improved voice cloning capabilities with greater diversity and accuracy
• Multilingual Support: New experimental support for Chinese, Japanese, and Korean languages

Speech Demo

Installation

pip install outetts --upgrade

Important:

• For GGUF support, install llama-cpp-python manually. Installation Guide
• For EXL2 support, install exllamav2 manually. Installation Guide

Usage

Quick Start: Basic Full Example

import outetts

# Configure the model
model_config = outetts.HFModelConfig_v1(
    model_path="OuteAI/OuteTTS-0.2-500M",
    language="en",  # Supported languages in v0.2: en, zh, ja, ko
)

# Initialize the interface
interface = outetts.InterfaceHF(model_version="0.2", cfg=model_config)

# Print available default speakers
interface.print_default_speakers()

# Load a default speaker
speaker = interface.load_default_speaker(name="male_1")

# Generate speech
output = interface.generate(
    text="Speech synthesis is the artificial production of human speech.",
    temperature=0.1,
    repetition_penalty=1.1,
    max_length=4096,

    # Optional: Use a speaker profile for consistent voice characteristics
    # Without a speaker profile, the model will generate a voice with random characteristics
    speaker=speaker,
)

# Save the generated speech to a file
output.save("output.wav")

# Optional: Play the generated audio
# output.play()

Backend-Specific Configuration

Hugging Face Transformers

import outetts

model_config = outetts.HFModelConfig_v1(
    model_path="OuteAI/OuteTTS-0.2-500M",
    language="en",  # Supported languages in v0.2: en, zh, ja, ko
)

interface = outetts.InterfaceHF(model_version="0.2", cfg=model_config)

GGUF (llama-cpp-python)

import outetts

model_config = outetts.GGUFModelConfig_v1(
    model_path="local/path/to/model.gguf",
    language="en", # Supported languages in v0.2: en, zh, ja, ko
    n_gpu_layers=0,
)

interface = outetts.InterfaceGGUF(model_version="0.2", cfg=model_config)

ExLlamaV2

import outetts

model_config = outetts.EXL2ModelConfig_v1(
    model_path="local/path/to/model",
    language="en", # Supported languages in v0.2: en, zh, ja, ko
)

interface = outetts.InterfaceEXL2(model_version="0.2", cfg=model_config)

Speaker Creation and Management

Creating a Speaker

You can create a speaker profile for voice cloning, which is compatible across all backends.

speaker = interface.create_speaker(
    audio_path="path/to/audio/file.wav",

    # If transcript is not provided, it will be automatically transcribed using Whisper
    transcript=None,            # Set to None to use Whisper for transcription

    whisper_model="turbo",      # Optional: specify Whisper model (default: "turbo")
    whisper_device=None,        # Optional: specify device for Whisper (default: None)
)

Saving and Loading Speaker Profiles

Speaker profiles can be saved and loaded across all supported backends.

# Save speaker profile
interface.save_speaker(speaker, "speaker.json")

# Load speaker profile
speaker = interface.load_speaker("speaker.json")

Default Speaker Initialization

OuteTTS includes a set of default speaker profiles. Use them directly:

# Print available default speakers
interface.print_default_speakers()
# Load a default speaker
speaker = interface.load_default_speaker(name="male_1")

Text-to-Speech Generation

The generation process is consistent across all backends.

output = interface.generate(
    text="Speech synthesis is the artificial production of human speech.",
    temperature=0.1,
    repetition_penalty=1.1,
    max_length=4096,
    speaker=speaker, # Optional: speaker profile
)

output.save("output.wav")
# Optional: Play the audio
# output.play()

Custom Backend Configuration

You can initialize custom backend configurations for specific needs.

Example with Flash Attention for Hugging Face Transformers

model_config = outetts.HFModelConfig_v1(
    model_path="OuteAI/OuteTTS-0.2-500M",
    language="en",
    dtype=torch.bfloat16,
    additional_model_config={
        'attn_implementation': "flash_attention_2"
    }
)

Speaker Profile Recommendations

To achieve the best results when creating a speaker profile, consider the following recommendations:

1. Audio Clip Duration:

   • Use an audio clip of around 10-15 seconds.
   • This duration provides sufficient data for the model to learn the speaker's characteristics while keeping the input manageable. The model's context length is 4096 tokens, allowing it to generate around 54 seconds of audio in total. However, when a speaker profile is included, this capacity is reduced proportionally to the length of the speaker's audio clip.

2. Audio Quality:

   • Ensure the audio is clear and noise-free. Background noise or distortions can reduce the model's ability to extract accurate voice features.

3. Accurate Transcription:

   • Provide a highly accurate transcription of the audio clip. Mismatches between the audio and transcription can lead to suboptimal results.

4. Speaker Familiarity:

   • The model performs best with voices that are similar to those seen during training. Using a voice that is significantly different from typical training samples (e.g., unique accents, rare vocal characteristics) might result in inaccurate replication.
   • In such cases, you may need to fine-tune the model specifically on your target speaker's voice to achieve a better representation.

5. Parameter Adjustments:

   • Adjust parameters like temperature in the generate function to refine the expressive quality and consistency of the synthesized voice.

Model Specifications

• Base Model: Qwen-2.5-0.5B
• Parameter Count: 500M
• Language Support:
  • Primary: English
  • Experimental: Chinese, Japanese, Korean
• License: CC BY NC 4.0

Training Datasets

• Emilia-Dataset (CC BY NC 4.0)
• LibriTTS-R (CC BY 4.0)
• Multilingual LibriSpeech (MLS) (CC BY 4.0)

Credits & References

• WavTokenizer
• CTC Forced Alignment
• Qwen-2.5-0.5B