NIMDeploymentConfig

class nemo_microservices.types.NIMDeploymentConfig(*args: Any, **kwargs: Any)

Bases: BaseModel

gpu: int

The number of GPUs needed for a deployment.

image_name: str

The name of the Docker image.

image_tag: str

The tag of the Docker image.

additional_envs: Dict[str, str] | None = None

Additional environment variables to pass to the deployment.

disable_lora_support: bool | None = None

When true, prevents setting default values for NIM_PEFT_SOURCE, NIM_PEFT_REFRESH_INTERVAL and related service-level environment variables. This will load the NIM without support for LoRAs. This is required for some NIMs which do not support LoRAs. This feature is experimental and may change in future versions.

Type:

EXPERIMENTAL

namespace: str | None = None

The Kubernetes namespace of the deployment.

pvc_size: str | None = None

The size of the PVC volume.