Skip to search

LoRAAdapter

inference.llmkube.dev / v1alpha1

apiVersion: inference.llmkube.dev/v1alpha1 kind: LoRAAdapter metadata: name: example
View raw schema
apiVersion string
APIVersion defines the versioned schema of this representation of an object. Servers should convert recognized schemas to the latest internal value, and may reject unrecognized values. More info: https://git.k8s.io/community/contributors/devel/sig-architecture/api-conventions.md#resources
kind string
Kind is a string value representing the REST resource this object represents. Servers may infer this from the endpoint the client submits requests to. Cannot be updated. In CamelCase. More info: https://git.k8s.io/community/contributors/devel/sig-architecture/api-conventions.md#types-kinds
metadata object
spec object required
spec defines the desired state of LoRAAdapter
inferenceServiceRef object required
InferenceServiceRef names the InferenceService this adapter should be loaded into. The reconciler resolves it; if it points at a non-sglang runtime, Available=False with reason RuntimeMismatch.
name string required
Name is the name of the target InferenceService. Must exist in Namespace (or be omitted to default to this resource's namespace).
namespace string
Namespace is the namespace of the target InferenceService. When omitted, defaults to the LoRAAdapter's namespace. Cross-namespace references are allowed; the controller only requires that the operator has RBAC to GET the referenced InferenceService.
name string required
Name is the SGLang-side adapter handle. Operators quote this in inference requests to pick the adapter. Must be unique within the target InferenceService's loaded set; the SGLang HTTP API rejects duplicates with 409.
pattern: ^[A-Za-z0-9][A-Za-z0-9._-]*$
minLength: 1
maxLength: 63
path string required
Path is the path on disk inside the SGLang container where the adapter weights are mounted. SGLang reads from this path when /load_lora_adapter is invoked, so it must be reachable from the inference pod (typically via a PVC exposed in InferenceService.spec.extraVolumes) at the path the operator chose. The controller does not auto-mount.
minLength: 1
status object
status defines the observed state of LoRAAdapter
conditions []object
conditions represent the current state of the LoRAAdapter resource. Standard condition types include: - "Available": the spec is well-formed and target InferenceService exists with runtime == sglang. - "Loaded": the adapter has been POSTed to SGLang and SGLang reported success (or, on teardown, acknowledged the unload). - "Error": a reconcile step failed; the message carries the underlying reason (runtime mismatch, HTTP failure, etc.). The status of each condition is one of True, False, or Unknown.
lastTransitionTime string required
lastTransitionTime is the last time the condition transitioned from one status to another. This should be when the underlying condition changed. If that is not known, then using the time when the API field changed is acceptable.
format: date-time
message string required
message is a human readable message indicating details about the transition. This may be an empty string.
maxLength: 32768
observedGeneration integer
observedGeneration represents the .metadata.generation that the condition was set based upon. For instance, if .metadata.generation is currently 12, but the .status.conditions[x].observedGeneration is 9, the condition is out of date with respect to the current state of the instance.
format: int64
minimum: 0
reason string required
reason contains a programmatic identifier indicating the reason for the condition's last transition. Producers of specific condition types may define expected values and meanings for this field, and whether the values are considered a guaranteed API. The value should be a CamelCase string. This field may not be empty.
pattern: ^[A-Za-z]([A-Za-z0-9_,:]*[A-Za-z0-9_])?$
minLength: 1
maxLength: 1024
status string required
status of the condition, one of True, False, Unknown.
enum: True, False, Unknown
type string required
type of condition in CamelCase or in foo.example.com/CamelCase.
pattern: ^([a-z0-9]([-a-z0-9]*[a-z0-9])?(\.[a-z0-9]([-a-z0-9]*[a-z0-9])?)*/)?(([A-Za-z0-9][-A-Za-z0-9_.]*)?[A-Za-z0-9])$
maxLength: 316
lastLoadedAt string
LastLoadedAt is when the adapter was last successfully loaded into SGLang. Re-loads (after a SGLang pod restart, for example) update this field.
format: date-time
loadedPath string
LoadedPath is the path SGLang has accepted for this adapter. Set after a successful POST /load_lora_adapter response. Empty before the first successful load.

No matches. Try .spec.inferenceServiceRef for an exact path