A Multi-Stage Multi-Codebook VQ-VAE Approach to High-Performance Neural TTS.
Haohan GuoFeng-Long XieFrank K. SoongXixin WuHelen MengPublished in: INTERSPEECH (2022)
Keyphrases
- multistage
- vector quantization
- vector quantizer
- image compression
- vector quantized
- production system
- dynamic programming
- text to speech
- single stage
- stochastic programming
- finite state vector quantization
- vector quantisation
- entropy constrained
- stochastic optimization
- attack detection
- neural network
- image representation
- codebook design
- lot sizing
- network architecture
- image classification
- visual words
- assembly systems
- codebook generation
- production line
- multiresolution
- image coding
- fractal image coding
- reinforcement learning