To create a sagemaker inference component (which enables you to utilize lora adapters), you must create an endpoint without a model, aka an empty production variant. In the aws SDK, production variants do not have a hard requirement for model name. In the pulumi provider for endpoint configurations, it does.

This means we can’t create empty variants, which means no empty endpoint to add our inference components to. If you try to add the inference component to an endpoint that already has a model you get

error: creating resource: creating resource (await): operation CREATE failed with "InvalidRequest": Invalid request provided: Inference Components are not supported in this Endpoint. Please make sure this endpoint can deploy inference components.

I’m going to create an issue for this, as it is actively blocking my work