88Z AI, 13 August — Xiaohongshu's dots team and Shanghai Jiao Tong University's X-LANCE Lab jointly open-sourced dots.tts, a 2-billion-parameter, fully continuous latent space autoregressive TTS base model. The model does not rely on discrete acoustic tokens, achieves SOTA on the zero-shot voice cloning benchmark Seed-TTS-Eval, and supports timbre fine-tuning and speech editing.
At a glance
- dots.tts is a 2-billion-parameter fully continuous latent space autoregressive model that does not rely on discrete acoustic tokens
- It achieved SOTA on the Seed-TTS-Eval zero-shot benchmark and supports timbre fine-tuning and speech editing
The update
The dots team at Xiaohongshu, in collaboration with the X-LANCE Lab at Shanghai Jiao Tong University, has officially open-sourced dots.tts, a speech synthesis base model. Featuring 2 billion parameters, the model utilizes a fully continuous latent space autoregressive architecture. Unlike traditional approaches, dots.tts does not rely on discrete acoustic tokens, enabling smoother generation. It achieved state-of-the-art performance on the Seed-TTS-Eval benchmark for zero-shot voice cloning. Furthermore, the model offers high flexibility, supporting timbre fine-tuning and direct speech editing, providing developers with a powerful foundational tool.
Sources & editorial note
This article is an AI-assisted synthesis by 88Z AI based on the public sources below. Product claims and availability are attributed to their publishers. It does not represent independent testing or first-hand reporting.
- 机器之心Source published
Source dates above refer to the original announcements; the date under the headline is this article’s publication on 88Z AI.