diff --git a/doc/README.md b/doc/README.md index 175b732..65e6ba7 100644 --- a/doc/README.md +++ b/doc/README.md @@ -6,6 +6,7 @@ FADI Documentation * [Reverse proxy](REVERSEPROXY.md) - Traefik reverse proxy configuration * [Security](SECURITY.md) - SSL setup * [TSimulus](TSIMULUS.md) - how to simulate sensors and generate realistic data with [TSimulus](https://github.com/cetic/TSimulus) +* [Machine learning models management](SELDON.md) - how to package and score machine learning models using [Seldon Core](https://www.seldon.io/tech/products/core/) * [Sample self-hosted infrastructure](RANCHER_PROXMOX.md) - How to install FADI on a self hosted infrastructure using * [Proxmox](https://www.proxmox.com/en/) as a self-hosted private cloud (IaaS) provider. It provides virtual machines for the various Kubernetes nodes. * [Rancher](https://rancher.com/what-is-rancher/what-rancher-adds-to-kubernetes/) to manage (install, provision, maintain, upgrade, ...) several Kubernetes clusters, e.g. when needing several environments on various IaaS providers or several well separated tenant installations, or doing airgapped installations on premises. diff --git a/doc/SELDON.md b/doc/SELDON.md new file mode 100644 index 0000000..ac9d32d --- /dev/null +++ b/doc/SELDON.md @@ -0,0 +1,98 @@ +Manage machine learning models with Seldon Core +========== + +* [Install Seldon Core service](#install-seldon-core-service) +* [Deploy your model](#deploy-your-model) + * [1. Package your model](#1-package-your-model) + * [2. Create your inference graph](#2-create-your-inference-graph) + * [3. Deploy the model to the Kubernetes cluster](#3-deploy-the-model-to-the-kubernetes-cluster) + +

+ + Seldon Core Logo + +

+ +[Seldon Core](https://www.seldon.io/tech/products/core/) is an open source platform for deploying machine learning models on a Kubernetes cluster. It extends Kubernetes with **its own custom resource `SeldonDeployment`** where you can define your runtime inference graph made up of models and other components that Seldon will manage. + +## Install Seldon Core service + +To deploy the Seldon Core service inside your FADI installation, set `seldon-core-operator.enabled` option to `true` in your FADI `values.yaml` configuration file and reapply the chart: + +```yaml +seldon-core-operator: + enabled: true + usageMetrics: + enabled: false +``` + +## Deploy your model + +### 1. Package your model + +To allow your component (model, router etc.) to be managed by Seldon Core it needs to be built into a **Docker container** and to expose the appropriate [service microservice APIs over REST or gRPC](https://docs.seldon.io/projects/seldon-core/en/latest/reference/apis/internal-api.html). + +To wrap your model follow the [official Seldon instructions](https://docs.seldon.io/projects/seldon-core/en/v1.1.0/python/index.html). + +NB: currently only Python is ready for production use, but other languages ([Java, R, Go, ...](https://docs.seldon.io/projects/seldon-core/en/latest/wrappers/language_wrappers.html)) are compatible. + +### 2. Create your inference graph + +Seldon Core extends Kubernetes with its own custom resource `SeldonDeployment` where you can define your runtime [inference graph](https://docs.seldon.io/projects/seldon-core/en/latest/graph/inference-graph.html) made up of models and other components that Seldon will manage. + +A `SeldonDeployment` is a JSON or YAML file that allows you to define your graph of component images and the resources each of those images will need to run (using a Kubernetes PodTemplateSpec). Below is a minimal example for a single model, in YAML: + +```yaml +apiVersion: machinelearning.seldon.io/v1alpha2 +kind: SeldonDeployment +metadata: + name: seldon-model +spec: + name: test-deployment + predictors: + - componentSpecs: + - spec: + containers: + - name: classifier + image: seldonio/mock_classifier:1.0 + graph: + children: [] + endpoint: + type: REST + name: classifier + type: MODEL + name: example + replicas: 1 +``` + +[ref](https://docs.seldon.io/projects/seldon-core/en/v1.1.0/graph/inference-graph.html) + +The key components are: + +* A list of **`predictors`**, each with a specification for the number of replicas. + * Each predictor defines a graph and its set of deployments. Having multiple predictors is useful when you want to split traffic between a main graph and a [canary](https://martinfowler.com/bliki/CanaryRelease.html), or for other production rollout scenarios. +* For each predictor, a **list of `componentSpecs`**. Each `componentSpec` is a Kubernetes `PodTemplateSpec` that Seldon will build into a Kubernetes Deployment. Place here the images from your graph and their requirements, e.g. `Volumes`, `ImagePullSecrets`, Resources Requests, etc. +* A **`graph`** specification that describes how the components are joined together. + +To understand the inference graph definition in detail see the [Seldon Deployment Reference Types + reference](https://docs.seldon.io/projects/seldon-core/en/latest/reference/seldon-deployment.html) + +### 3. Deploy the model to the Kubernetes cluster + +Once the inference graph is created as a JSON or YAML Seldon Deployment resource, you can deploy it to the Kubernetes cluster: + +```bash +kubectl apply -f my_deployment.yaml +``` + +To delete ( or manage ) your `SeldonDeployment` you can use kubectl for the custom resource `SeldonDeployment`, for example to see if there are any models deployed: + +```bash +kubectl get seldondeployment +``` + +To delete the model `seldon-model`: + +```bash +kubectl delete seldondeployment seldon-model +``` diff --git a/doc/images/logos/seldon_logo.jpg b/doc/images/logos/seldon_logo.jpg new file mode 100644 index 0000000..ed325e3 Binary files /dev/null and b/doc/images/logos/seldon_logo.jpg differ