maple.cmd.cli.serve.serve_policy

maple.cmd.cli.serve.serve_policy(name: str = <typer.models.ArgumentInfo object>, port: int = <typer.models.OptionInfo object>, device: str = <typer.models.OptionInfo object>, host_port: int | None = <typer.models.OptionInfo object>, model_load_kwargs: str = <typer.models.OptionInfo object>) → None

Serve a policy model in a container.

Requests the daemon to start a policy container with the specified model. The policy is loaded onto the specified device and made available via HTTP API for inference requests.

Parameters:

• name – Policy specification string (name or name:version).

• port – Daemon port number.

• device – Device to load policy on (e.g., ‘cuda:0’, ‘cpu’).

• host_port – Optional specific port to bind the policy container to.

• model_load_kwargs – Model-specific loading parameters.