KonfAI KonfAI KonfAI
/

Getting started

  • Installation
  • Quickstart

Core concepts

  • Core concepts
    • Configuration model
    • Datasets and groups
    • Model graph and output naming
    • Execution flow
    • KonfAI Apps
  • Configuration reference
    • Training configuration
    • Prediction configuration
    • Evaluation configuration
    • Common configuration patterns

Guides

  • How-to guides
    • Training workflows
    • Prediction workflows
    • Evaluation workflows
    • Using custom models, transforms, augmentations, and losses
    • Using KonfAI Apps
    • Remote server and client mode
    • Docker usage
    • Testing and validation

Reference

  • Reference
    • CLI reference
    • Environment variables
    • App server HTTP API
    • API reference
      • Public API overview
      • Workflows API
      • Configuration API
      • Data API
      • Models API
      • Apps API
      • Extension points
      • Full module reference
        • konfai package

Project

  • Examples
    • Synthesis example
    • Segmentation example
  • Troubleshooting
  • Contributing
  • Architecture and internals
  1. KonfAI /
  2. Configuration reference

Configuration reference¶

These pages document the canonical YAML structure used by KonfAI’s three main workflows.

They focus on the fields that are stable and clearly visible in the codebase and the shipped examples. For built-in models, transforms, and metrics, the exact available parameters still depend on the selected classpath.

  • Training configuration
  • Prediction configuration
  • Evaluation configuration
  • Common configuration patterns
Previous
KonfAI Apps
Next
Training configuration

2026, Valentin Boussot

Made with Sphinx and Shibuya theme.