Skip to content

Seldon

tl; dr; A combinator data component that installs Seldon-Core, an ML serving framework.

Introduction

Seldon-Core is an open-source model serving and monitoring framework. It allows you to deploy your ML models so that they can be consumed by users over consistent REST APIs. In addition, other Seldon tools allow you to monitor your model in production.

Learn more about Seldon-Core in the repo.

Test Drive

The fastest way to get started is to use the test drive functionality provided by TestFaster. Click on the "Launch Test Drive" button below (opens a new window).

💻 Launch Test Drive 💻

Launch Jupyter

Once the component has launched, click on the Jupyter link. Seldon-Core does not have a UI by default. You will use Jupyter to interact with Seldon-Core via its Kubernetes API.

Example Notebook

Once inside Jupyter, browse to the demo notebook, which comes from the official quick start guide. Follow the instructions in the notebook to deploy a pre-trained model.

Usage

Prerequisites

Start by preparing your Kubernetes cluster using one of the infrastructure components or use your own cluster.

Component Usage

module "seldon" {
  source  = "combinator-ml/seldon/k8s"
  # Optional settings go here
}

See the full configuration options below.

Requirements

Name Version
helm >= 2.0.0
kubectl >= 1.7.0
kubernetes >= 2.0.0

Providers

Name Version
helm >= 2.0.0
kubectl >= 1.7.0
kubernetes >= 2.0.0

Modules

Name Source Version
istio combinator-ml/istio/k8s 0.0.1

Resources

Name
helm_release
kubectl_file_documents
kubectl_manifest
kubernetes_namespace

Inputs

Name Description Type Default Required
enable_example_seldon_deployment Enable an example seldon deployment bool true no
enable_seldon_gateway Create an istio gateway for seldon bool true no
seldon_core_operator_namespace (Optional) The namespace to install the minio operator into. Defaults to minio-operator string "seldon-system" no
seldon_core_values (Optional) List of values in raw yaml to pass to helm. list(string) [] no

Outputs

No output.