Comments (7)
你好,stargan跟cyclegan的整体框架是差不多的,但是这份代码里确实用的是stargan,而且代码很多参考了stargan原始论文放出的代码。我没有仔细调参,初步试验转换结果还行。谢谢!
from stargan-voice-conversion.
你好,模型是基于stargan的,但不是stargan-vc论文里的模型结构,这个论文你有了解吗,我看你在生成器和判别器最后一层都加了分类器,是什么作用?
from stargan-voice-conversion.
stargan-vc 里面用一个Discriminator和一个Classifier,这里实现的也有一个Discriminator和一个Classifier,只是我让他们共享大部分参数,只有在靠近输出端才分出两支来。
看model.py的78、79行
self.conv_dis = nn.Conv2d(curr_dim, 1, kernel_size=(kernel_size_0, kernel_size_1), stride=1, padding=0, bias=False) # padding should be 0
self.conv_clf_spks = nn.Conv2d(curr_dim, num_speakers, kernel_size=(kernel_size_0, kernel_size_1), stride=1, padding=0, bias=False) # for num_speaker
self.conv_dis
这个是输出Real/Fake。
self.conv_clf_spks
这个是分类用的,它的输出维度是num_speakers
。
当然你也可以跟stargan-vc原paper一样,不共享参数。
生成器后面没有加分类器。
from stargan-voice-conversion.
恩,分类器共享判别器的参数了,不知单独设置分类器和共享判别器参数的分类器对模型的影响如何。
最好能够放一些转换后的样本上来听一下。
from stargan-voice-conversion.
Hello,你好,我已经star了~
from stargan-voice-conversion.
from stargan-voice-conversion.
stargan-vc 里面用一个Discriminator和一个Classifier,这里实现的也有一个Discriminator和一个Classifier,只是我让他们共享大部分参数,只有在靠近输出端才分出两支来。 看model.py的78、79行
self.conv_dis = nn.Conv2d(curr_dim, 1, kernel_size=(kernel_size_0, kernel_size_1), stride=1, padding=0, bias=False) # padding should be 0 self.conv_clf_spks = nn.Conv2d(curr_dim, num_speakers, kernel_size=(kernel_size_0, kernel_size_1), stride=1, padding=0, bias=False) # for num_speaker
self.conv_dis
这个是输出Real/Fake。self.conv_clf_spks
这个是分类用的,它的输出维度是num_speakers
。 当然你也可以跟stargan-vc原paper一样,不共享参数。 生成器后面没有加分类器。
请问,加分类器C为什么可以优化D和G?就是多考虑了一种损失吗
from stargan-voice-conversion.
Related Issues (20)
- I cannot run the code. HOT 24
- Do you have file ./models\200000-G.ckpt ? I want to download it. Thank you
- preprocessing.py possible sox issue HOT 4
- Id mapping loss HOT 1
- Loss function meanings HOT 4
- Suggestions for documentation
- Number of Mel-cpestral coefficients (MCEPs)
- Why g_loss is lack of g_loss_identity
- not find gated cnn
- How to fine-tune StarGAN-VC model?
- Error in training with more than 4 speakers
- D/loss_real: -0.0000
- Inference time HOT 3
- Can implementation of the author share 200000 iteration model for comparative study? HOT 5
- A question about the adversarial loss.
- Are there any requirements for training datasets? HOT 2
- Python 3.5 HOT 1
- run Convert.py wrong HOT 1
- How should I take it?Thank you! HOT 2
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
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.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from stargan-voice-conversion.