Dragon-FM generates speech autoregressively over two-second chunks while using flow matching inside each chunk, achieving fast synthesis at 12.5 discrete audio tokens per second.
istftnet: Fast and lightweight mel-spectrogram vocoder incorporating inverse short-time fourier transform
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.SD 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Next Tokens Denoising for Speech Synthesis
Dragon-FM generates speech autoregressively over two-second chunks while using flow matching inside each chunk, achieving fast synthesis at 12.5 discrete audio tokens per second.