---
language:
- ja
license: mit
tags:
- audio
- speech-synthesis
- VITS
- gaming
- Arknights
datasets:
- custom
widget:
- text: 你好，我是艾雅法拉。欢迎回到罗德岛。
---

# GPT-SoVITS: Arknights Eyjafjalla Voice Synthesis

![image/png](https://cdn-uploads.huggingface.co/production/uploads/655a156d9d249b4ab35f85db/QTYPIcP9Fbb8y6XpozOUH.png)

<audio controls src="https://cdn-uploads.huggingface.co/production/uploads/655a156d9d249b4ab35f85db/x9LBg0kcG5nJeJb4aZm1G.wav"></audio>
「最近新しいカフェがオープンしたの知ってる？とってもおしゃれで、ケーキも美味しいんだって。週末にでも、時間があったら一緒に行ってみない？楽しい時間を過ごせると思うよ。」

## Model Description

This model is a state-of-the-art voice synthesis tool based on the [GPT-SoVITS](https://github.com/RVC-Boss/GPT-SoVITS) architecture, specifically trained to replicate the voice of Eyjafjalla, a popular character from the game Arknights. The model achieves exceptional quality and high fidelity in voice generation, closely mimicking Eyjafjalla's unique vocal characteristics.

## How to Use

This model is designed for easy integration and use through a GPT-SoVITS webUI. To generate voice clips, simply load the provided `GPT_weights` and `SoVITS_weights` files into the web interface and input your desired text. The model will then infer and produce a voice clip that closely resembles Eyjafjalla's voice from Arknights.

## Training Data

The model was trained using a custom dataset compiled just from game dialogues.

## Applications

**Attention: Commercial use is prohibited, proceed at your own risk.**

This model can be used for various applications, including but not limited to:
- Game development
- Fan-made content creation
- Virtual assistants
- Educational tools

## License

This model is released under the MIT License. For more information, please see the LICENSE file.