Notebooks — seldon-core documentation (original) (raw)
Seldon Core Setup¶
Prepackaged Inference Server Examples¶
- Deploy a Scikit-learn Model Binary
- Deploy a Tensorflow Exported Model
- MLflow Pre-packaged Model Server A/B Test
- MLflow v2 Protocol End to End Workflow (Incubating)
- Deploy a XGBoost Model Binary
- Deploy Pre-packaged Model Server with Cluster's MinIO
- Custom Pre-packaged LightGBM Server
Python Language Wrapper Examples¶
- SKLearn Spacy NLP
- SKLearn Iris Classifier
- Sagemaker SKLearn Example
- TFserving MNIST
- Statsmodels Holt-Winter's time-series model
- Runtime Metrics & Tags
- Triton GPT2 Example
Specialised Framework Examples¶
Incubating Projects Examples¶
- Kubeflow Seldon E2E Pipeline
- H2O Java MoJo
- Outlier Detection with Combiner
- Stream Processing with KNative Eventing
- Kafka CIFAR10
- Kafka SpaCy SKlearn NLP
- Kafka KEDA Autoscaling
- CPP Wrapper Simple Single File
- Advanced CPP Buildsystem Override
- Environment Variables
Cloud-Specific Examples¶
- AWS EKS Tensorflow Deep MNIST
- Azure AKS Tensorflow Deep MNIST
- GKE with GPU Tensorflow Deep MNIST
- Alibaba Cloud Tensorflow Deep MNIST
- Triton GPT2 Example Azure
- Setup Azure Kubernetes Infrastructure
- Create Storage Account for training data
- Install Kubernetes Blob CSI Driver
Advanced Machine Learning Monitoring¶
- Real Time Monitoring of Statistical Metrics
- Explainer for Iris model with Poetry-defined Environment
- Deploy Iris Model
- Train Explainer
- Save and deploy Explainer
- Test Deployed explainer
- Model Explainer V2 protocol Example (Incubating)
- Outlier Detection on CIFAR10
- Outlier Detector for Cifar10 model with Poetry-defined Environment
- Train Outlier Detector
Batch Processing with Seldon Core¶
- Batch Processing with Argo Workflows and S3 / Minio
- Batch Processing with Argo Workflows and HDFS
- Batch Processing with Kubeflow Pipelines
MLOps: Scaling and Monitoring and Observability¶
- Autoscaling Example
- KEDA Autoscaling example
- Request Payload Logging with ELK
- Custom Metrics with Grafana & Prometheus
- Distributed Tracing with Jaeger
- CI / CD with Jenkins Classic
- CI / CD with Jenkins X
- Replica control
Production Configurations and Integrations¶
- Example Helm Deployments
- Max gRPC Message Size
- Deploy Multiple Seldon Core Operators
- Protocol Examples
- Configurable timeouts
- Custom Protobuf Data Example
- Disruption Budgets Example
AB Tests and Progressive Rollouts¶
- Istio AB Test
- Ambassador AB Test
- Seldon/Iter8 - Progressive AB Test with Single Seldon Deployment
- Seldon/Iter8 - Progressive AB Test with Multiple Seldon Deployments
Complex Graph Examples¶
- Chainer MNIST Model Deployment
- Test using Minikube
- Custom pre-processors with the V2 Protocol
- graph-examples
Ingress¶
Infrastructure¶
Benchmarking and Load Tests¶
- Service Orchestrator Overhead
- Tensorflow Benchmark
- Argo Workflows Benchmarking
- Python Serialization Cost Benchmark
- KMP_AFFINITY Benchmarking Example
- Kafka Payload Logging