Whether To Pretrain DNN or Not?: An Empirical Analysis for Voice Conversion
会议：2019 interspeech 作者：Nirmesh J. Shah 单位：DA-IICT, Gandhinagar, India abstract 在数据有限的情况下，vc model会过拟合，因此大部分的方法是pre-training再finetune。作者希望探究通过不需要预训练的方法同时避免过拟合。 1. introduction 深度学习中，预训练的方法逐渐被better act
相关文章
- 1. 2019trans--Sequence-to-Sequence Acoustic Modeling for Voice Conversion
- 2. How NOT To Evaluate Your Dialogue System: An Empirical Study of Unsupervised Evaluation Metrics for
- 3. FeatureNMS-- the rule whether to add p from P to D or not is adjusted
- 4. 2018ins--Wavelet Analysis of Speaker Dependent and Independent Prosody for Voice Conversion
- 5. An Empirical Study of Real-world Polymorphic Code Injection Attacks
- 6. Check whether the number is **"Prime number"** or not
- 7. StarGAN-VC2: Rethinking Conditional Methods for StarGAN-Based Voice Conversion
- 8. 2016 ICME:Phonetic posteriorgrams for many-to-one voice conversion without parallel data training
- 9. To forgive or not
- 10. ONE-SHOT VOICE CONVERSION USING STAR-GAN
- 更多相关文章...
- • Scala for循环 - Scala教程
- • SQL NOT NULL 约束 - SQL 教程
- • Kotlin学习（二）基本类型
- • Github 简明教程
相关标签/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。
欢迎关注本站公众号,获取更多信息
相关文章
- 1. 2019trans--Sequence-to-Sequence Acoustic Modeling for Voice Conversion
- 2. How NOT To Evaluate Your Dialogue System: An Empirical Study of Unsupervised Evaluation Metrics for
- 3. FeatureNMS-- the rule whether to add p from P to D or not is adjusted
- 4. 2018ins--Wavelet Analysis of Speaker Dependent and Independent Prosody for Voice Conversion
- 5. An Empirical Study of Real-world Polymorphic Code Injection Attacks
- 6. Check whether the number is **"Prime number"** or not
- 7. StarGAN-VC2: Rethinking Conditional Methods for StarGAN-Based Voice Conversion
- 8. 2016 ICME:Phonetic posteriorgrams for many-to-one voice conversion without parallel data training
- 9. To forgive or not
- 10. ONE-SHOT VOICE CONVERSION USING STAR-GAN