seldon-single-model

Version: 0.2.0

Chart to deploy a machine learning model in Seldon Core.

Usage

使用本chart,首先要添加 seldonio Helm 仓库:

helm repo add seldonio https://storage.googleapis.com/seldon-charts
helm repo update

一旦完成,你可以按照如下使用预估图模板:

helm template $MY_MODEL_NAME seldonio/seldon-single-model --namespace $MODELS_NAMESPACE

注意你也可以直接部署预估图到集群中: 使用:

helm install $MY_MODEL_NAME seldonio/seldon-single-model --namespace $MODELS_NAMESPACE

Homepage: https://github.com/SeldonIO/seldon-core

源码

设置值

类型

默认值

描述

annotations

object

{}

Annotations applied to the deployment

apiVersion

string

"machinelearning.seldon.io/v1"

Version of the SeldonDeployment CRD

hpa.enabled

bool

false

Whether to add an HPA spec to the deployment

hpa.maxReplicas

int

5

Maximum number of replicas for HPA

hpa.metrics

list

[{"resource":{"name":"cpu","targetAverageUtilization":10},"type":"Resource"}]

Metrics that autoscaler should check

hpa.minReplicas

int

1

Minimum number of replicas for HPA

labels

object

{}

Labels applied to the deployment

model.env

object

{"LOG_LEVEL":"INFO"}

Environment variables injected into the model’s container

model.envSecretRefName

string

""

The model secret name for enviroment variables

model.image

string

""

Docker image used by the model

model.implementation

string

""

Implementation of Prepackaged Model Server

model.logger.enabled

bool

false

model.logger.url

string

""

model.mlflow.xtype

string

""

model.resources

object

{"requests":{"memory":"1Mi"}}

Resource requests and limits for the model’s container

model.uri

string

""

Model’s URI for prepackaged model server

protocol

string

"seldon"

replicas

int

1

Number of replicas for the predictor