| --- |
| license: apache-2.0 |
| tags: |
| - sherpa-onnx |
| - speech-to-text |
| - text-to-speech |
| - voice-activity-detection |
| --- |
| |
| # RunAnywhere models |
|
|
| Speech, vision and voice-activity assets used by the RunAnywhere example apps. |
|
|
| These were previously fetched from GitHub release pages. They live here so every |
| model the apps download comes from one place, with the same host, auth and |
| resume behaviour as the rest of the catalog. |
|
|
| Files are grouped by capability. |
|
|
| | Folder | Contents | |
| |---|---| |
| | `stt/` | Whisper tiny, base and small, packaged for sherpa-onnx | |
| | `tts/` | Piper VITS voices (Alba, Amy, Lessac, Ryan) and Supertonic 3 int8 | |
| | `vad/` | Silero VAD | |
| | `vlm/` | SmolVLM 500M instruct | |
|
|
| Both `.tar.gz` and `.tar.bz2` builds of Whisper tiny and the Lessac voice are |
| kept, because the desktop and mobile apps expect different archive formats. |
|
|
| ## Where these came from |
|
|
| Nothing here was trained or modified by RunAnywhere. Each file is a byte-for-byte |
| copy of an upstream release. |
|
|
| - Whisper and Piper packaging, and Supertonic 3, come from |
| [k2-fsa/sherpa-onnx](https://github.com/k2-fsa/sherpa-onnx) releases (Apache-2.0). |
| - Silero VAD comes from [snakers4/silero-vad](https://github.com/snakers4/silero-vad) (MIT). |
| - Whisper is [OpenAI's](https://github.com/openai/whisper) (MIT), Piper is |
| [Rhasspy's](https://github.com/rhasspy/piper) (MIT), and SmolVLM is |
| [Hugging Face's](https://huggingface.co/HuggingFaceTB) (Apache-2.0). |
|
|
| Original licenses apply. Please credit the upstream projects rather than this mirror. |
|
|
