kalefranz/seldon-core

Machine Learning Deployment for Kubernetes

Seldon Core

Branch	Status
master
release-0.2
release-0.1

Seldon Core is an open source platform for deploying machine learning models on Kubernetes.

• Goals
• Quick Start
• Example Components
• Integrations
• Install
• Deployment guide
• Reference
• Article/Blogs/Videos
• Community
• Developer
• Latest Seldon Images
• Usage Reporting

Goals

Machine learning deployment has many challenges. Seldon Core intends to help with these challenges. Its high level goals are:

• Allow data scientists to create models using any machine learning toolkit or programming language. We plan to initially cover the tools/languages below:
  • Python based models including
    • Tensorflow models
    • Sklearn models
  • Spark models
  • H2O models
  • R models
• Expose machine learning models via REST and gRPC automatically when deployed for easy integration into business apps that need predictions.
• Allow complex runtime inference graphs to be deployed as microservices. These graphs can be composed of:
  • Models - runtime inference executable for machine learning models
  • Routers - route API requests to sub-graphs. Examples: AB Tests, Multi-Armed Bandits.
  • Combiners - combine the responses from sub-graphs. Examples: ensembles of models
  • Transformers - transform request or responses. Example: transform feature requests.
• Handle full lifecycle management of the deployed model:
  • Updating the runtime graph with no downtime
  • Scaling
  • Monitoring
  • Security

Prerequisites

A Kubernetes Cluster. Kubernetes can be deployed into many environments, both on cloud and on-premise.

Quick Start

Read the overview to using seldon-core.

• Jupyter notebooks showing examples:
  • Seldon Core Deployments using Helm
  • Seldon Core Deployments using Ksonnet

Example Components

Seldon-core allows various types of components to be built and plugged into the runtime prediction graph. These include models, routers, transformers and combiners. Some example components that are available as part of the project are:

• Models : example that illustrate simple machine learning models to help you build your own integrations

  • Python
    • Tensorflow MNIST Classifier
    • Keras MNIST Classifier
    • Scikit-learn MNIST Classifier
    • Scikit-learn Iris Classifier
  • R
    • R MNIST Classifier
    • R Iris Classifier
  • Java
    • H2O Classifier
  • NodeJS
    • Tensorflow MNIST Classifier
  • ONNX
    • ResNet ONNX Classifier using Intel nGraph
  • PMML
    • PySpark MNIST Classifier

• routers

  • Epsilon-greedy multi-armed bandits for real time optimization of models

• transformers

  • Mahalanobis distance outlier detection. Example usage can be found in the Advanced graphs notebook

Integrations

• kubeflow
  • Seldon-core can be installed as part of the kubeflow project. A detailed end-to-end example provides a complete workflow for training various models and deploying them using seldon-core.
• IBM's Fabric for Deep Learning
  • Seldon-core can be used to serve deep learning models trained using FfDL.
    • Train and deploy a Tensorflow MNIST classififer using FfDL and Seldon.
    • Train and deploy a PyTorch MNIST classififer using FfDL and Seldon.
• Istio and Seldon
  • Canary deployemts using Istio and Seldon..
• NVIDIA TensorRT and DL Inference Server
• Tensorflow Serving
• Intel OpenVINO
  • A Helm chart for easy integration and an example notebook using OpenVINO to serve imagenet model within Seldon Core.

Install

Follow the install guide for details on ways to install seldon onto your Kubernetes cluster.

Deployment Guide

Three steps:

1. Wrap your runtime prediction model.
   • We provide easy to use wrappers for python, R, Java and NodeJS.
   • We have tools to test your wrapped components.
2. Define your runtime inference graph in a seldon deployment custom resource.
3. Deploy the graph.

Advanced Tutorials

• Advanced graphs showing the various types of runtime prediction graphs that can be built.
• Handling large gRPC messages. Showing how you can add annotations to increase the gRPC max message size.
• Handling REST timeouts. Showing how you can add annotations to set the REST (and gRPC) timeouts.

Reference

• Prediction API
  • Proto Buffer Definitions
  • Open API Definitions
• Seldon Deployment Custom Resource
• Analytics

Articles/Blogs/Videos

• Kubecon Europe 2018 - Serving Machine Learning Models at Scale with Kubeflow and Seldon
• Polyaxon, Argo and Seldon for model training, package and deployment in Kubernetes
• Manage ML Deployments Like A Boss: Deploy Your First AB Test With Sklearn, Kubernetes and Seldon-core using Only Your Web Browser & Google Cloud
• Using PyTorch 1.0 and ONNX with Fabric for Deep Learning
• AI on Kubernetes - O'Reilly Tutorial
• Scalable Data Science - The State of DevOps/MLOps in 2018
• Istio Weekly Community Meeting - Seldon-core with Istio
• Openshift Commons ML SIG - Openshift S2I Helping ML Deployment with Seldon-Core
• Overview of Openshift source-to-image use in Seldon-Core
• IBM Framework for Deep Learning and Seldon-Core
• CartPole game by Reinforcement Learning, a journey from training to inference

Release Highlights

• 0.2.3 Release Highlights

Testing

• Benchmarking seldon-core

Configuration

• Annotation based configuration.
• Notes for running in production.
• Helm configuration
• ksonnet configuration

Community

• Slack Channel

Developer

• CHANGELOG
• Developer Guide

Latest Seldon Images

Description	Image URL	Stable Version	Development
Seldon Operator	seldonio/cluster-manager	0.2.3	0.2.4-SNAPSHOT
Seldon Service Orchestrator	seldonio/engine	0.2.3	0.2.4-SNAPSHOT
Seldon API Gateway	seldonio/apife	0.2.3	0.2.4-SNAPSHOT
Seldon Python 3 Wrapper for S2I	seldonio/seldon-core-s2i-python3	0.2	0.3-SNAPSHOT
Seldon Python 2 Wrapper for S2I	seldonio/seldon-core-s2i-python2	0.2	0.3-SNAPSHOT
Seldon Python ONNX Wrapper for S2I	seldonio/seldon-core-s2i-python3-ngraph-onnx	0.1
Seldon Core Python Wrapper	seldonio/core-python-wrapper	0.7
Seldon Java Build Wrapper for S2I	seldonio/seldon-core-s2i-java-build	0.1
Seldon Java Runtime Wrapper for S2I	seldonio/seldon-core-s2i-java-runtime	0.1
Seldon R Wrapper for S2I	seldonio/seldon-core-s2i-r	0.1
Seldon NodeJS Wrapper for S2I	seldonio/seldon-core-s2i-nodejs	0.1	0.2-SNAPSHOT
Seldon Tensorflow Serving proxy	seldonio/tfserving-proxy	0.1
Seldon NVIDIA inference server proxy	seldonio/nvidia-inference-server-proxy	0.1

Java Packages

Description	Package	Version
Seldon Core Wrapper	seldon-core-wrapper	0.1.2
Seldon Core JPMML	seldon-core-jpmml	0.0.1

Usage Reporting

Tools that help the development of Seldon Core from anonymous usage.

• Usage Reporting with Spartakus