Publish & deploy

Once you have a GGUF you’re happy with, publishing it is two steps:

Upload to Hugging Face. The CLI creates or reuses a Hugging Face repo and uploads the GGUF there. This step needs an HF_TOKEN.
Assign to an app. “Deploying” a model means assigning it to an app in the GresIQ model catalog (assign_model(app_id, model_id)). The end app then fetches that assignment through the Onde SDK’s load_assigned_model and downloads the GGUF itself — the CLI never pushes bytes to the device directly.

The CLI can only assign models that already exist in the GresIQ catalog, so publish uploads the model first and registers it in the catalog as part of the same flow.

See GresIQ & deployment for how the catalog and app assignment fit together, and the SDK section for how an app picks up an assigned model at runtime.