VoxMel open-sourced: wavelet-factorized AR mel-spectrogram TTS models
_akhaliq · x · 2026-11-12
VoxDance-Lab has open-sourced VoxMel on Hugging Face, a zero-shot text-to-speech model family using wavelet-factorized autoregressive mel-spectrogram modeling. Two checkpoints are available: VoxMel-0.3B and VoxMel-0.6B.
More from Multimodal
- Ant's Ling-3.2-flash: 560B/25B active multimodal model beats Opus 5 on vision benchmarks — Med1_Ai · 2026-11-12
- Blogger shares an AI-generated comedy clip that's actually funny — FinanceYF5 · 2026-11-12
- Hailuo AI shows off homework maps made with its H3Image model — Hailuo_AI · 2026-11-12
- vLLM team open-sources OpenDirector, an agent-driven AI video creation workflow — vllm_project · 2026-11-12
- vLLM-Omni brings Day-0 support for MiniMax H3-Image: native 4K in 0.87s — vllm_project · 2026-11-12
- Spotify to let Premium users AI-remix songs with royalties for original artists from 2027 — nikola_mr64990 · 2026-11-12