marswh12312313 / aistella

基于 BERT-VITS2 的月见光沅的 Ai 音色模型。

AiStella

基于 BERT-VITS2 的月见光沅 Ai 音色模型。

Summary

The Ai timbre model of Stella based on BERT-VITS2. Welcome to the AiStella repository! AiStella is an innovative AI model that replicates the unique speaking style of Stella using the power of BERT for understanding context and VITS2 for speech synthesis.

Introduction

AiStella combines the contextual understanding abilities of BERT with the state-of-the-art voice synthesis capabilities of VITS2 to create a highly realistic voice model. BERT, which stands for Bidirectional Encoder Representations from Transformers, provides a deep understanding of language nuances, which is then used by VITS2, a cutting-edge neural network for generating human-like speech, to produce a voice that not only sounds like Stella but also captures her speech mannerisms.

Models

• AiStella v1.0
• ... Coming Soon!

Notice

We are thrilled to share that AiStella is an ongoing project, and we are diligently working on the next big release. Version 2.0 is around the corner, and it promises to bring a plethora of enhancements and innovative features that will take Stella's voice synthesis to the next level. Keep an eye out for updates, and thank you for your support of AiStella.

For any queries or contributions, feel free to reach out by opening an issue or a pull request. Your feedback is invaluable for the progress of AiStella.

Thank you for exploring AiStella, where the future of personalized voice AI comes to life.

Exciting Features Await in Version 2.0 – Stay Tuned!

---

Please be aware that the usage of AiStella should comply with all relevant laws and regulations. The model data is copyrighted, and any application of the AiStella model must respect the data rights and legal boundaries.