End-to-End Emotional Speech Synthesis Using Style Tokens and Semi-Supervised Training.
Peng-Fei WuZhen-Hua LingLi-Juan LiuYuan JiangHong-Chuan WuLi-Rong DaiPublished in: CoRR (2019)
Keyphrases
- end to end
- speech synthesis
- supervised training
- speech recognition
- text to speech
- unsupervised learning
- conditional random fields
- training phase
- recognition process
- neural nets
- supervised learning
- dependency parsing
- admission control
- training algorithm
- congestion control
- graphical models
- neural network
- text localization and recognition
- application layer
- higher order
- training data
- image sequences
- computer vision