StarGAN-VC2: Rethinking Conditional Methods for StarGAN-Based Voice Conversion
該論文鏈接原文地址 發表於Interspeech 2019 總體概括： 1.使用了很多損失函數 對抗 分類（區別人） 身份（進一步對目標的映射）循環一致性損失 ，這些損失都來源於STARGA 以及Cycle-GAN。 2.一對多 其它很多是一對一 該方法是一對多 此處對應損失函數分類和身份 循環一致性損失 3.非平行數據 同時論文中表明爲何使用非平行數據，因爲平行數據很多還需要人工的
相關文章
- 1. StarGAN-VC2: Rethinking Conditional Methods for StarGAN-Based Voice Conversion
- 2. 2019trans--Sequence-to-Sequence Acoustic Modeling for Voice Conversion
- 3. ONE-SHOT VOICE CONVERSION USING STAR-GAN
- 4. SINGING VOICE CONVERSION WITH NON-PARALLEL DATA
- 5. Whether To Pretrain DNN or Not?: An Empirical Analysis for Voice Conversion
- 6. Voice Conversion using Convolutional Neural Networks 翻譯
- 7. 2018ins--Wavelet Analysis of Speaker Dependent and Independent Prosody for Voice Conversion
- 8. 2016 ICME:Phonetic posteriorgrams for many-to-one voice conversion without parallel data training
- 9. Rethinking Class-Balanced Methods for Long-Tailed Visual Recognition論文翻譯
- 10. StarGAN-VC： non-parallel many-to-many voice conversion with StaGAN
- 更多相關文章...
- • Swift for 循環 - Swift 教程
- • Scala for循環 - Scala教程
- • RxJava操作符（七）Conditional and Boolean
- • PHP開發工具
相關標籤/搜索
每日一句
-
每一个你不满意的现在，都有一个你没有努力的曾经。
歡迎關注本站公眾號,獲取更多信息
相關文章
- 1. StarGAN-VC2: Rethinking Conditional Methods for StarGAN-Based Voice Conversion
- 2. 2019trans--Sequence-to-Sequence Acoustic Modeling for Voice Conversion
- 3. ONE-SHOT VOICE CONVERSION USING STAR-GAN
- 4. SINGING VOICE CONVERSION WITH NON-PARALLEL DATA
- 5. Whether To Pretrain DNN or Not?: An Empirical Analysis for Voice Conversion
- 6. Voice Conversion using Convolutional Neural Networks 翻譯
- 7. 2018ins--Wavelet Analysis of Speaker Dependent and Independent Prosody for Voice Conversion
- 8. 2016 ICME:Phonetic posteriorgrams for many-to-one voice conversion without parallel data training
- 9. Rethinking Class-Balanced Methods for Long-Tailed Visual Recognition論文翻譯
- 10. StarGAN-VC： non-parallel many-to-many voice conversion with StaGAN