Xiaohongshu Open-Sources dots.tts: 2B Continuous AR Speech Model
机器之心 · wechat · 2026-08-13
Xiaohongshu's dots team, collaborating with Shanghai Jiao Tong University, has open-sourced dots.tts, a 2B parameter end-to-end text-to-speech foundation model. Departing from the mainstream discrete token approach, dots.tts performs autoregressive generation directly in a continuous latent space, preserving richer acoustic details like tone and prosody.
On popular zero-shot voice cloning benchmarks such as Seed-TTS-Eval, dots.tts achieves SOTA in both average content accuracy and speaker similarity. The model natively supports streaming output, achieving an ultra-low audio first-package latency of 54.4ms for duplex dialogue systems. The release includes six checkpoints covering base models, self-corrected alignment, and multi-step/single-step distillation, alongside fully open training and fine-tuning code.
Related event: Xiaohongshu Open-Sources 2B Parameter TTS Model dots.tts(2 posts)→
More from Multimodal
- Flux 3 Klien first impressions: occasional body horror like SD3, otherwise solid — rionix88 · 2026-10-16
- XGRIDS builds world's largest 3DGS town: 23.5 km² and 6.2 billion Gaussian Splats — bilawalsidhu · 2026-10-16
- IMMERSION: AI film compilation links 12 shorts with one continuous camera move — AI_Andrew · 2026-10-16
- A song with vocals, lyrics and 3D head animation created entirely by Opus 5.5 — repligate · 2026-10-16
- Alibaba open-sources CommerceGen-Studio for commercial-grade e-commerce images — alibabagroup · 2026-10-16
- EVA-Cap: Fudan and OPPO align video captioning with atomic audiovisual events via GRPO — Fudan-University · 2026-10-16