TransformerState<T>

Models & Types · AiDotNet.Interfaces

Represents the serializable state of a fitted time series transformer.

For Beginners

This class allows you to save a fitted transformer to a file and reload it later without needing to re-fit. This is useful for:

  • Production deployments where you train once and deploy the fitted model
  • Sharing trained transformers between team members
  • Versioning and reproducibility of your feature engineering pipeline

Constructors

Constructor Summary
TransformerState()

Properties

Property Summary
Dictionary<string, object> Options { get; set; } The serialized options used to configure the transformer.
Dictionary<string, object> Parameters { get; set; } Transformer-specific parameters and learned values.
IncrementalState<T> IncrementalState { get; set; } The incremental state, if initialized.
bool IsFitted { get; set; } Whether the transformer has been fitted.
int InputFeatureCount { get; set; } The number of input features.
int OutputFeatureCount { get; set; } The number of output features.
int Version { get; set; } The version of the serialization format.
int[] WindowSizes { get; set; } The window sizes used by the transformer.
string TransformerType { get; set; } The type name of the transformer.
string[] FeatureNames { get; set; } The generated feature names.
string[] InputFeatureNames { get; set; } The input feature names.