zhenye234 / comospeech Goto Github PK
View Code? Open in Web Editor NEWCoMoSpeech: One-Step Speech and Singing Voice Synthesis via Consistency Model
License: MIT License
CoMoSpeech: One-Step Speech and Singing Voice Synthesis via Consistency Model
License: MIT License
Hello, thank you for sharing this project. I was hoping to test it out for inference and fine-tuning but could not see any pretrained models available to do so, would it be possible to share an english model?
Is text-to-speech the only feature available right now?
RT.
The paper mentions using CoMoSpeech for the SVS task and briefly describes feature extraction, but I can't find that feature extraction or summing with the phoneme features in the code. Is that planned for release in the future?
Thanks!
Does singing voice work on English? songs or not?
Hi, I replace zh pinyin phonemes and finish the network for multi speakers (without feed spk_emb to GradLogPEstimator2d), when I train teacher model on custom datasets with 17 speakers (each speakers 6000 wavs) about 108 epochs, the teacher model generate audio still contain distinct electric sound, is that normal?
Moreover, I change network to feed spk_emb to GradLogPEstimator2d and find the network can quickly reconstruct the timbre of different speakers (only about 2 epoch), but still contain distinct electric sound, how can I work out for it? or just wait more epochs?
Looking forward for your kind reply
如题
作者你好,读论文看Eq.(7)还是没懂是如何推导出来的,drift coefficient为f(t)=t,diffusion coefficient为g(t)=1,那么\sigma(t)=t,然后再结合Eq.(4),请问是带入到Eq.(3)这个ODE吗?我自己推了下还是没得到Eq.(7)。不知道是我哪里理解有误?谢谢!
因为本文的教师模型为grad-TTS,该模型仅针对语音合成,所以想问一下如果想要复现SVS部分:
1.需要用什么模型作为教师模型?参数如何设置?
2.已经获得opencpop数据集,需要做哪些预处理?修改哪些训练参数?
3.声码器可以使用como-SVC的声码器(m4singer-hifigan)吗?如果不行,需要修改哪些参数?
非常感谢您的解答:)
A declarative, efficient, and flexible JavaScript library for building user interfaces.
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google ❤️ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.