Giter Club home page Giter Club logo

Comments (14)

hongwen-sun avatar hongwen-sun commented on May 20, 2024 10

@MaxMax2016 你好,我做了个对比的实验,使用不同ppg训练声学模型的情况。训练集的loss如下:
ppg_resultpng

whisper:medium
contentvec: checkpoint_best_legacy_500
hubertsoft: hubert-soft-0d54a1f4.pt

from lora-svc.

MaxMax2016 avatar MaxMax2016 commented on May 20, 2024 2

@hongwen-sun 赞,有疑惑然后去验证,是个搞科研的好手!

from lora-svc.

hongwen-sun avatar hongwen-sun commented on May 20, 2024 2

@MaxMax2016 您好,可以加一下微信吗,可以做一些交流?
[email protected] 如果可以的话,可以发邮件给我,我加您

from lora-svc.

nshmyrev avatar nshmyrev commented on May 20, 2024 1

Btw, there is argumentation behind this that whisper was actually trained on songs. It is very good at speech recognition of songs compared to many other speech models.

from lora-svc.

MaxMax2016 avatar MaxMax2016 commented on May 20, 2024

whisper is from openai .

from lora-svc.

loboere avatar loboere commented on May 20, 2024

mm.. does that mean it's better?

from lora-svc.

MaxMax2016 avatar MaxMax2016 commented on May 20, 2024

from now on, whisper is the best audio encoder for svc.这个项目的目的就是为了去证明这一点。

from lora-svc.

hongwen-sun avatar hongwen-sun commented on May 20, 2024

from now on, whisper is the best audio encoder for svc.这个项目的目的就是为了去证明这一点。

您有对比几版相同模型不同content提取方式的结果吗?比如hubert,whisper,content vec,或者类似so vit svc的第九层的content vec就用您现在的方案。
另外我去看了so vit svc的方案,用了vits的框架。相对来说,您的方案只用了声码器部分结构,相对简洁但是模型参数量和模型能力是足够的吗?
我粗浅理解是从效果上看,您的方案很难达到一个完整vits达到的效果。

from lora-svc.

MaxMax2016 avatar MaxMax2016 commented on May 20, 2024

@hongwen-sun 你我的目标不同,您期望的是一个音质效果好的SVC;我是在做技术方案的研究,比如whisper用于歌声转换的有效性,还有说话人自适应;我需要一个高效的方案来验证我的一些想法。

from lora-svc.

MaxMax2016 avatar MaxMax2016 commented on May 20, 2024

此外,我个人比较崇尚简洁。您如何看待 ‘奥卡姆剃刀原理’ 呢?

from lora-svc.

hongwen-sun avatar hongwen-sun commented on May 20, 2024

我感觉简洁也可以达到比较好的效果,并不冲突,效果的差距通过block结构的简单修改和参数量的增加完全可以弥补。
另外,如果验证有效性,我理解还是需要对比不同content的提取方法在你这套框架下的效果,不然很难证明whisper是best encoder,只能说是一个可行的方案,这个基本上你实验前就知道了。
当然,您这个仓库确实是个很好的项目,我也有参考到,非常感谢。

from lora-svc.

hongwen-sun avatar hongwen-sun commented on May 20, 2024

说白了,我本来是想白嫖您的结论,看看如果您做了对比,哪个好一点

from lora-svc.

MaxMax2016 avatar MaxMax2016 commented on May 20, 2024

whisper是目前为止能获取到的效果最好的多语言ASR模型,它的好坏由他的识别率就可以直接体现,它通过海量的多语言数据训练得到,这是其他开源模型无法媲美的。如果您要做语音转换可能hubert等其他自监督模型比较合适,它们包含了其他非语言的信息、如语气&情感等,也许还有性别信息、以及可能的泄漏的音色;

from lora-svc.

hongwen-sun avatar hongwen-sun commented on May 20, 2024

好的,非常感谢。
对于转换而言,whisper也不一定就比不过自监督模型,更可能的情况是这些模型都有自身的缺点,需要针对vc任务有专门的设计。

from lora-svc.

Related Issues (20)

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google ❤️ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.