Login / Signup
Mixer-TTS: Non-Autoregressive, Fast and Compact Text-to-Speech Model Conditioned on Language Model Embeddings.
Oktai Tatanov
Stanislav Beliaev
Boris Ginsburg
Published in:
ICASSP (2022)
Keyphrases
</>
autoregressive
text to speech
language model
probabilistic model
speech recognition
random fields
maximum entropy
similarity measure
relevance model
moving average
dependency structure