VoxRT/linnet-en-35m
archived · low-signalmodel
Linnet is a lightweight on-device text-to-speech model: an autoregressive flow language model that emits one continuous 48-dimensional latent per 80 ms step (two 24-dimensional codec frames at 25 Hz) with a single flow evaluation per step, followed by a Vocos-style neural codec decoder that turns… License: linnet-weights-license-1.0.
- Org
- VoxRT
- Released
- 2026-11-04
- Status
- archived · low-signal
- License
- linnet-weights-license-1.0 (restricted)
- Task
- text-to-speech · audio
- Signal
- 7 HF likes · 298 downloads
- Archived
- 2026-11-07 (low-signal: aged out below threshold; stats as last seen)
Links
First published here 2026-11-04 · updated 2026-11-15 · markdown · JSON