DiFlow-TTS: Discrete Flow Matching with Factorized Speech Tokens for Low-Latency Zero-Shot Text-To-Speech
Paper • 2509.09631 • Published
DiFlow-TTS: Compact and Low-Latency Zero-Shot Text-to-Speech with Discrete Flow Matching
DiFlow-TTS is trained on 470 hours of the LibriTTS dataset, which consists of predominantly neutral speech. As a result, it may not perform well on prompts with strong emotional expression.
Download DiFlow-TTS checkpoint, and place it as follows:
root/
└── ckpts/
└── diflow-tts.ckpt