README.md

Packages:

• serving.kubeflow.org/v1alpha2

serving.kubeflow.org/v1alpha2

Package v1alpha2 contains API Schema definitions for the serving v1alpha2 API group

Resource Types:

AlibiExplainerSpec

(Appears on: ExplainerSpec)

AlibiExplainerSpec defines the arguments for configuring an Alibi Explanation Server

Field	Description
type AlibiExplainerType	The type of Alibi explainer
storageUri string	The location of a trained explanation model
runtimeVersion string	Alibi docker image version which defaults to latest release
resources Kubernetes core/v1.ResourceRequirements	Defaults to requests and limits of 1CPU, 2Gb MEM.
config map[string]string	Inline custom parameter settings for explainer

AlibiExplainerType (string alias)

(Appears on: AlibiExplainerSpec)

Batcher

(Appears on: DeploymentSpec)

Batcher provides optional payload batcher for all endpoints

Field	Description
maxBatchSize int	(Optional) MaxBatchSize of batcher service
maxLatency int	(Optional) MaxLatency of batcher service
timeout int	(Optional) Timeout of batcher service

ComponentStatusMap

EndpointStatusMap defines the observed state of InferenceService endpoints

CustomSpec

(Appears on: ExplainerSpec, PredictorSpec, TransformerSpec)

CustomSpec provides a hook for arbitrary container configuration.

Field	Description
container Kubernetes core/v1.Container

DeploymentSpec

(Appears on: ExplainerSpec, PredictorSpec, TransformerSpec)

DeploymentSpec defines the configuration for a given InferenceService service component

Field	Description
serviceAccountName string	(Optional) ServiceAccountName is the name of the ServiceAccount to use to run the service
minReplicas int	(Optional) Minimum number of replicas which defaults to 1, when minReplicas = 0 pods scale down to 0 in case of no traffic
maxReplicas int	(Optional) This is the up bound for autoscaler to scale to
parallelism int	Parallelism specifies how many requests can be processed concurrently, this sets the hard limit of the container concurrency(https://knative.dev/docs/serving/autoscaling/concurrency).
logger Logger	(Optional) Activate request/response logging
batcher Batcher	(Optional) Activate request batching

EndpointSpec

(Appears on: InferenceServiceSpec)

Field	Description
predictor PredictorSpec	Predictor defines the model serving spec
explainer ExplainerSpec	(Optional) Explainer defines the model explanation service spec, explainer service calls to predictor or transformer if it is specified.
transformer TransformerSpec	(Optional) Transformer defines the pre/post processing before and after the predictor call, transformer service calls to predictor service.

Explainer

ExplainerConfig

(Appears on: ExplainersConfig)

Field	Description
image string
defaultImageVersion string

ExplainerSpec

(Appears on: EndpointSpec)

ExplainerSpec defines the arguments for a model explanation server, The following fields follow a “1-of” semantic. Users must specify exactly one spec.

Field	Description
alibi AlibiExplainerSpec	Spec for alibi explainer
custom CustomSpec	Spec for a custom explainer
DeploymentSpec DeploymentSpec	(Members of DeploymentSpec are embedded into this type.)

ExplainersConfig

(Appears on: InferenceServicesConfig)

Field	Description
alibi ExplainerConfig

InferenceService

InferenceService is the Schema for the services API

Field	Description
metadata Kubernetes meta/v1.ObjectMeta	Refer to the Kubernetes API documentation for the fields of the metadata field.
spec InferenceServiceSpec	default EndpointSpec Default defines default InferenceService endpoints canary EndpointSpec (Optional) Canary defines alternate endpoints to route a percentage of traffic. canaryTrafficPercent int (Optional) CanaryTrafficPercent defines the percentage of traffic going to canary InferenceService endpoints
status InferenceServiceStatus

InferenceServiceSpec

(Appears on: InferenceService)

InferenceServiceSpec defines the desired state of InferenceService

Field	Description
default EndpointSpec	Default defines default InferenceService endpoints
canary EndpointSpec	(Optional) Canary defines alternate endpoints to route a percentage of traffic.
canaryTrafficPercent int	(Optional) CanaryTrafficPercent defines the percentage of traffic going to canary InferenceService endpoints

InferenceServiceState (string alias)

InferenceState describes the Readiness of the InferenceService

InferenceServiceStatus

(Appears on: InferenceService)

InferenceServiceStatus defines the observed state of InferenceService

Field	Description
Status knative.dev/pkg/apis/duck/v1beta1.Status	(Members of Status are embedded into this type.)
url string	URL of the InferenceService
traffic int	Traffic percentage that goes to default services
canaryTraffic int	Traffic percentage that goes to canary services
default map[github.com/kubeflow/kfserving/pkg/constants.InferenceServiceComponent]github.com/kubeflow/kfserving/pkg/apis/serving/v1alpha2.StatusConfigurationSpec	Statuses for the default endpoints of the InferenceService
canary map[github.com/kubeflow/kfserving/pkg/constants.InferenceServiceComponent]github.com/kubeflow/kfserving/pkg/apis/serving/v1alpha2.StatusConfigurationSpec	Statuses for the canary endpoints of the InferenceService
address knative.dev/pkg/apis/duck/v1beta1.Addressable	Addressable URL for eventing

InferenceServicesConfig

Field	Description
transformers TransformersConfig
predictors PredictorsConfig
explainers ExplainersConfig

Logger

(Appears on: DeploymentSpec)

Logger provides optional payload logging for all endpoints

Field	Description
url string	(Optional) URL to send request logging CloudEvents
mode LoggerMode	What payloads to log: [all, request, response]

LoggerMode (string alias)

(Appears on: Logger)

ONNXSpec

(Appears on: PredictorSpec)

ONNXSpec defines arguments for configuring ONNX model serving.

Field	Description
storageUri string	The URI of the exported onnx model(model.onnx)
runtimeVersion string	ONNXRuntime docker image versions, default version can be set in the inferenceservice configmap
resources Kubernetes core/v1.ResourceRequirements	Defaults to requests and limits of 1CPU, 2Gb MEM.

Predictor

PredictorConfig

(Appears on: PredictorsConfig)

Field	Description
image string
defaultImageVersion string
defaultGpuImageVersion string

PredictorSpec

(Appears on: EndpointSpec)

PredictorSpec defines the configuration for a predictor, The following fields follow a “1-of” semantic. Users must specify exactly one spec.

Field	Description
custom CustomSpec	Spec for a custom predictor
tensorflow TensorflowSpec	Spec for Tensorflow Serving (https://github.com/tensorflow/serving)
triton TritonSpec	Spec for Triton Inference Server (https://github.com/NVIDIA/triton-inference-server)
xgboost XGBoostSpec	Spec for XGBoost predictor
sklearn SKLearnSpec	Spec for SKLearn predictor
onnx ONNXSpec	Spec for ONNX runtime (https://github.com/microsoft/onnxruntime)
pytorch PyTorchSpec	Spec for PyTorch predictor
DeploymentSpec DeploymentSpec	(Members of DeploymentSpec are embedded into this type.)

PredictorsConfig

(Appears on: InferenceServicesConfig)

Field	Description
tensorflow PredictorConfig
triton PredictorConfig
xgboost PredictorConfig
sklearn PredictorConfig
pytorch PredictorConfig
onnx PredictorConfig

PyTorchSpec

(Appears on: PredictorSpec)

PyTorchSpec defines arguments for configuring PyTorch model serving.

Field	Description
storageUri string	The URI of the trained model which contains model.pt
modelClassName string	Defaults PyTorch model class name to ‘PyTorchModel’
runtimeVersion string	PyTorch KFServer docker image version which defaults to latest release
resources Kubernetes core/v1.ResourceRequirements	Defaults to requests and limits of 1CPU, 2Gb MEM.

SKLearnSpec

(Appears on: PredictorSpec)

SKLearnSpec defines arguments for configuring SKLearn model serving.

Field	Description
storageUri string	The URI of the trained model which contains model.pickle, model.pkl or model.joblib
runtimeVersion string	SKLearn KFServer docker image version which defaults to latest release
resources Kubernetes core/v1.ResourceRequirements	Defaults to requests and limits of 1CPU, 2Gb MEM.

StatusConfigurationSpec

StatusConfigurationSpec describes the state of the configuration receiving traffic.

Field	Description
name string	Latest revision name that is in ready state
host string	Host name of the service

TensorflowSpec

(Appears on: PredictorSpec)

TensorflowSpec defines arguments for configuring Tensorflow model serving.

Field	Description
storageUri string	The URI for the saved model(https://www.tensorflow.org/tutorials/keras/save_and_load)
runtimeVersion string	TFServing docker image version(https://hub.docker.com/r/tensorflow/serving), default version can be set in the inferenceservice configmap.
resources Kubernetes core/v1.ResourceRequirements	Defaults to requests and limits of 1CPU, 2Gb MEM.

Transformer

Transformer interface is implemented by all Transformers

TransformerConfig

(Appears on: TransformersConfig)

Field	Description
image string
defaultImageVersion string

TransformerSpec

(Appears on: EndpointSpec)

TransformerSpec defines transformer service for pre/post processing

Field	Description
custom CustomSpec	Spec for a custom transformer
DeploymentSpec DeploymentSpec	(Members of DeploymentSpec are embedded into this type.)

TransformersConfig

(Appears on: InferenceServicesConfig)

Field	Description
feast TransformerConfig

TritonSpec

(Appears on: PredictorSpec)

TritonSpec defines arguments for configuring Triton Inference Server.

Field	Description
storageUri string	The URI for the trained model repository(https://docs.nvidia.com/deeplearning/triton-inference-server/master-user-guide/docs/model_repository.html)
runtimeVersion string	Triton Inference Server docker image version, default version can be set in the inferenceservice configmap
resources Kubernetes core/v1.ResourceRequirements	Defaults to requests and limits of 1CPU, 2Gb MEM.

VirtualServiceStatus

VirtualServiceStatus captures the status of the virtual service

Field	Description
URL string
CanaryWeight int
DefaultWeight int
address knative.dev/pkg/apis/duck/v1beta1.Addressable	(Optional) Address holds the information needed for a Route to be the target of an event.
Status knative.dev/pkg/apis/duck/v1beta1.Status

XGBoostSpec

(Appears on: PredictorSpec)

XGBoostSpec defines arguments for configuring XGBoost model serving.

Field	Description
storageUri string	The URI of the trained model which contains model.bst
nthread int	Number of thread to be used by XGBoost
runtimeVersion string	XGBoost KFServer docker image version which defaults to latest release
resources Kubernetes core/v1.ResourceRequirements	Defaults to requests and limits of 1CPU, 2Gb MEM.

---

Generated with gen-crd-api-reference-docs on git commit d7f65bc.