Statements (29)
| Predicate | Object |
|---|---|
| gptkbp:instanceOf |
gptkb:model
neural vocoder |
| gptkbp:application |
speech synthesis
text-to-speech |
| gptkbp:architecture |
flow-based generative model
|
| gptkbp:author |
gptkb:Rafael_Valle
gptkb:Bryan_Catanzaro gptkb:Ryan_Prenger |
| gptkbp:citation |
1000+
|
| gptkbp:developedBy |
gptkb:NVIDIA
|
| gptkbp:firstPublished |
2018
|
| https://www.w3.org/2000/01/rdf-schema#label |
WaveGlow
|
| gptkbp:input |
mel-spectrogram
|
| gptkbp:language |
gptkb:Python
|
| gptkbp:license |
gptkb:BSD-3-Clause
|
| gptkbp:notableFeature |
single network for fast, high-quality audio synthesis
does not require GANs does not require auto-regression |
| gptkbp:notablePublication |
gptkb:WaveGlow:_A_Flow-based_Generative_Network_for_Speech_Synthesis
https://arxiv.org/abs/1811.00002 |
| gptkbp:openSource |
true
|
| gptkbp:output |
audio waveform
|
| gptkbp:platform |
gptkb:PyTorch
|
| gptkbp:relatedTo |
gptkb:Glow
gptkb:WaveNet |
| gptkbp:repository |
https://github.com/NVIDIA/waveglow
|
| gptkbp:usedFor |
real-time speech synthesis
|
| gptkbp:bfsParent |
gptkb:NVIDIA_AI_Research
|
| gptkbp:bfsLayer |
6
|