christina-TTS-1.5

Christina TTS 1.5

This model features a custom voice inspired by the character Kurisu Makise from the visual novel 'Steins;Gate' — now in seven languages.

Version 1.5 extends Christina TTS with French, Spanish, Russian, German and Chinese, each with its own dedicated voice variant.

About the Voice

Important disclaimer: The voice actress in this model is not Asami Imai or any other official voice actor. This is a fan recreation using synthetic and real data, which explains any differences from the original character's voice.

The voice actress who lent her voice wishes to remain private. Thank you for your understanding.

Speakers

Seven voice variants, one per language:

Why Per-Language Variants?

Each speaker is anchored to voice references in its own language, so it produces natural intonation and rhythm there. christina and christina-jp carry over the exact voice identities from version 1.0. Using a variant outside its language works, but may carry an accent — which can be desirable for certain character portrayals.

Choose the variant that matches your output language for the most natural results.

Quickstart

Installation

pip install -U qwen-tts
# Optional: for optimized performance
pip install -U flash-attn --no-build-isolation

Quick Usage

import torch
import soundfile as sf
from qwen_tts import Qwen3TTSModel

model = Qwen3TTSModel.from_pretrained(
    "Loke-60000/christina-TTS-1.5",
    device_map="cuda:0",
    dtype=torch.bfloat16,
    attn_implementation="flash_attention_2",
)

# English
wavs_en, sr = model.generate_custom_voice(
    text="I finally managed to finish the experiment.",
    speaker="christina",
    language="English",
)
sf.write("christina_en.wav", wavs_en[0], sr)

# French
wavs_fr, sr = model.generate_custom_voice(
    text="J'ai enfin réussi à terminer l'expérience.",
    speaker="christina-fr",
    language="French",
)
sf.write("christina_fr.wav", wavs_fr[0], sr)

You can also inspect the available voices with model.get_supported_speakers().

CPU / Quantized Version

A 4-bit GGUF build for CPU-only inference (no Python, no GPU) is available at Loke-60000/christina-TTS-1.5-q4.

License & Usage

This model is free to download and use. However, no training data will be shared.

If you use this model in any project, please provide credit.