Wallaroo

Take your boldest algorithms live against production data faster, simpler, and at far lower cost.

Contents:

• Overview
• Single Sign On
• Models
  • Example
  • API Reference
• Models and Pipelines are Core SDK Abstractions
  • Convenience Style - single model
  • Convenience Style - chain a second model on the end
  • Convenience Style - replace a model
  • Replace a model in the middle of a pipeline
  • Explicit Style with all options
• Experiments
  • Random Split Experiment Example
  • Key Split Experiment Example
  • Shadow Deploy Experiment Example
  • API Reference
• Monitoring
  • Value Validation
  • Baseline Drift Detection
  • Alerting
• Model Insights
  • Workflow
  • Interactive Baseline Runs
  • Interactive Assay Runs
  • Default configuration
• Python Runtime
  • Uploading
  • Entrypoint
  • Tensor fields
  • State
  • Example Pipeline
  • Running inference on the pipeline
  • Under the hood
• Standalone Mode
  • Install Prerequisites
  • Install the Engine image
  • Use the SDK to generate models and configuration files
  • Launch standalone engine
  • Perform inference
  • Standalone engine HTTP API
  • Run integration tests
• Workspaces
  • Default Workspace
  • Create a Workspace
  • Add a user
  • Add an owner
  • Remove a user
  • Set Current Workspace
  • All new models and piplines will be uploaded to the current workspace
  • You can list workspaces and select from the returned array to switch workspaces
• Autoscaling
  • Example
• SDK
  • wallaroo
  • auth
  • client
  • checks
  • deployment_config
  • deployment
  • inference_result
  • logs
  • model_config
  • model
  • pipeline_variant
  • pipeline
  • standalone_client
• Jupyter
  • Jupyter Hub
  • Inside the Hub Environment
  • External Environment

Indices and tables

• Index

• Module Index

• Search Page