VersioningConfig

Options & Configuration · AiDotNet.Deployment.Configuration

Configuration for model versioning - managing multiple versions of the same model.

For Beginners

As you improve your AI model over time, you’ll have multiple versions. Versioning helps you manage these versions, allowing you to:

  • Keep track of which version is deployed
  • Roll back to a previous version if needed
  • Gradually transition users from old to new versions
  • Compare performance between versions

Version Format: Follows semantic versioning (e.g., “1.2.3”)

  • Major version: Breaking changes (1.0.0 → 2.0.0)
  • Minor version: New features, backwards compatible (1.0.0 → 1.1.0)
  • Patch version: Bug fixes (1.0.0 → 1.0.1)

You can also use “latest” to always get the newest version, or “stable” for the most reliable production version.

Constructors

Constructor Summary
VersioningConfig()

Properties

Property Summary
Dictionary<string, string> VersionMetadata { get; set; } Gets or sets the version metadata dictionary for storing additional version information.
bool AllowAutoUpgrade { get; set; } Gets or sets whether to allow automatic version upgrades (default: false).
bool AutoCleanup { get; set; } Gets or sets whether to automatically clean up old versions when MaxVersionHistory is exceeded (default: true).
bool Enabled { get; set; } Gets or sets whether versioning is enabled (default: true).
bool TrackVersionUsage { get; set; } Gets or sets whether to track which version is used for each inference (default: true).
int MaxVersionHistory { get; set; } Gets or sets the maximum number of versions to keep in history (default: 5).
int MaxVersionsPerModel { get; set; } Alias for MaxVersionHistory for more intuitive access.
string DefaultVersion { get; set; } Gets or sets the default version to use when none is specified (default: “latest”).