Properties (51)
| Predicate | Object |
|---|---|
| gptkbp:instanceOf |
deep learning model
|
| gptkbp:appliesTo |
audio generation
|
| gptkbp:basedOn |
autoregressive models
|
| gptkbp:completed |
state-of-the-art results
|
| gptkbp:developedBy |
DeepMind
|
| gptkbp:firstIntroduced |
2016
|
| gptkbp:has |
multiple layers
high sample quality |
| gptkbp:hasRelatedPatent |
music generation
|
| https://www.w3.org/2000/01/rdf-schema#label |
WaveNet
|
| gptkbp:improves |
naturalness of speech
|
| gptkbp:is |
computationally intensive
a significant advancement in AI a model that has been implemented in various frameworks a framework for audio synthesis a generative model a hierarchical model a model for sequential data a model that can be conditioned on text a model that can be fine-tuned a model that can be parallelized a model that can be scaled up a model that can produce diverse outputs a model that captures temporal dependencies a model that has been adopted in industry a model that has been cited extensively a model that has been open-sourced a model that has influenced other models. a model that has inspired further research a model that has led to new research directions a model that has received significant attention a model that learns from data a model that uses dilated convolutions a probabilistic model a type of neural network non-causal model a model that has been evaluated on various benchmarks a model that can be used in real-world applications |
| gptkbp:isFacilitatedBy |
complex audio signals
|
| gptkbp:isInfluencedBy |
WaveRNN
|
| gptkbp:isNear |
real-time system
|
| gptkbp:isPartOf |
traditional text-to-speech systems
|
| gptkbp:isTestedFor |
backpropagation
|
| gptkbp:isUsedFor |
voice cloning
|
| gptkbp:mayHave |
raw audio waveforms
human-like speech |
| gptkbp:powerOutput |
audio samples
|
| gptkbp:requires |
large datasets
|
| gptkbp:usedFor |
text-to-speech synthesis
|
| gptkbp:uses |
generative modeling
|
| gptkbp:utilizes |
convolutional neural networks
|