FastDiff: A Fast Conditional Diffusion Model for High-Quality Speech Synthesis.
Rongjie HuangMax W. Y. LamJun WangDan SuDong YuYi RenZhou ZhaoPublished in: CoRR (2022)
Keyphrases
- diffusion model
- speech synthesis
- high quality
- speech recognition
- text to speech
- anisotropic diffusion
- diffusion models
- vocal tract
- prosodic features
- information diffusion
- diffusion process
- laplace transform
- diffusion tensor
- image quality
- speech signal
- partial differential equations
- data mining
- hidden markov models
- high resolution
- magnetic resonance images
- flow field
- supply chain
- social networks