Text-to-Speech Models
ClariNet
Introduced by Ping et al. in ClariNet: Parallel Wave Generation in End-to-End Text-to-SpeechClariNet is an end-to-end text-to-speech architecture. Unlike previous TTS systems which use text-to-spectogram models with a separate waveform synthesizer (vocoder), ClariNet is a text-to-wave architecture that is fully convolutional and can be trained from scratch. In ClariNet, the WaveNet module is conditioned on the hidden states instead of the mel-spectogram. The architecture is otherwise based on Deep Voice 3.
Source: ClariNet: Parallel Wave Generation in End-to-End Text-to-Speech
Papers
| Paper | Code | Results | Date | Stars |
|---|
Tasks
| Task | Papers | Share |
|---|---|---|
| Speech Synthesis | 3 | 30.00% |
| Domain Adaptation | 2 | 20.00% |
| Unsupervised Domain Adaptation | 2 | 20.00% |
| Melody Extraction | 1 | 10.00% |
| Retrieval | 1 | 10.00% |
| Text-To-Speech Synthesis | 1 | 10.00% |
Bridge-net
Dense Connections
DV3 Attention Block
DV3 Convolution Block
L1 Regularization
Leaky ReLU
Normalizing Flows
ReLU
Softsign Activation
WaveNet
Weight Normalization
