简介 Intro
原神崩铁角色语音 BertVITS2 模型 V1 版是一款专为模拟游戏《原神》、《崩铁》角色语音而设计的人工智能语音合成模型。该模型基于先进的深度学习技术，结合了 BERT 的语言理解能力和 VITS 的语音合成技术，实现了高度自然、流畅的语音输出。它能够根据文本输入，生成与崩铁角色声音特征相匹配的语音，提供沉浸式的游戏体验。V1 版在语音的自然度、情感表达以及语音合成的准确性方面进行了优化，使得角色语音更加贴近原角色的声音特点，为玩家带来更加真实和愉悦的听觉享受。
The hoyoTTS Model is an artificial intelligence speech synthesis model designed to simulate the voice of characters in Genshin and Star Rail. The model is based on advanced deep learning technology, combining the language understanding capability of BERT and the speech synthesis technology of VITS. It realizes highly natural and smooth speech output. Based on the text input, it can generate voices that match the voice characteristics of the characters, providing an immersive gaming experience. v1 has been optimized in terms of voice naturalness, emotional expression, and the accuracy of voice synthesis, making the character voices closer to the original character's voice characteristics, and bringing players a more realistic and pleasurable auditory enjoyment.
在线体验 Demo
https://www.modelscope.cn/studios/Genius-Society/hoyoTTS
使用 Usage
from modelscope import snapshot_download
model_dir = snapshot_download('Genius-Society/hoyoTTS')
维护 Maintenance
git clone git@hf.co:Genius-Society/hoyoTTS
cd hoyoTTS
参考引用 Reference
[1] https://www.modelscope.cn/studios/erythrocyte/Bert-VITS2_Genshin_TTS