runntime-moonshine-streaming

Safetensors weights for Moonshine Streaming, repacked for runntime.

Upstream: https://huggingface.co/moonshine-ai
runntime: https://github.com/software-mansion/runntime

Layout

<size>/<precision>/model.safetensors
size precision
tiny f16, f32

tokenizer.json at the repo root is shared by every size.

Modifications

These are not the upstream files. Repacked by Software Mansion on 2026-09-14:

  • converted from the upstream checkpoint format to safetensors
  • cast to the precisions listed above to reduce download size
  • training-only tensors dropped; only the inference path is kept
  • laid out as <size>/<precision>/ for the runntime loader

No weight values were retrained or fine-tuned.

License

mit, inherited from Moonshine Streaming. Full text in LICENSE.

Downloads last month

-

Downloads are not tracked for this model. How to track
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for software-mansion/runntime-moonshine-streaming

Finetuned
(1)
this model