add-model

To scaffold a new KServe inference model server, you generate a model_server directory containing a model.py with standard load, preprocess, and predict methods, alongside Blubber and Docker Compose configurations for deployment.

Adding a new model requires updating pipeline/config.yaml, adding a blubber.yaml file for pipeline directories, modifying Docker Compose for mounted artifacts, and connecting external CI triggers through integration configuration files.

Yes, scaffolding a model server includes adding a unit test folder and connecting external CI triggers, ensuring the new KServe inference service is exercised and released reliably through automated pipelines.

Configuring Docker Compose for a KServe model server involves adding a service entry that exposes a port, mounts model artifacts, and sets the correct environment variables required to build and serve the inference model locally.

The required structure for a KServe model server code file is a model.py that subclasses kserve.Model and implements standard load, preprocess, and predict methods to ensure consistent inference behavior across services.

Yes, you can use this scaffolding process for any named model in a Lift Wing inference-services repository by providing a kebab-case name to generate all required code, pipeline publishing, and deployment configurations.