Module install_model 

omni-dev voice install-model — one-time fetch of model artefacts.

Supports two variants: whisper-tiny.en for the whisper-candle ASR backend, and speaker-wespeaker-en for the speaker-embedding runtime added in #805 / ADR-0034. Files land in the conventional install locations beneath ~/.omni-dev/voice/models/.

Bumps the model-download cost to install time rather than transcribe/ enrol time, so network failures surface explicitly when the user opts in to installing rather than silently on first use.

Structs

InstallModelCommand

Downloads the model files for a chosen variant into the conventional install location at ~/.omni-dev/voice/models/<variant-subdir>/ (or --dest to override).

Enums

Variant

Which model variant to install.