TOTEMOptions<T>

Options & Configuration · AiDotNet.Models.Options

Configuration options for TOTEM (TOkenized Time Series EMbeddings).

Inheritance: ModelOptions → RegressionOptions<T> → TimeSeriesRegressionOptions<T> → TOTEMOptions<T>

For Beginners

TOTEM bridges continuous time series and discrete tokens:

Vector Quantization (VQ-VAE): TOTEM maintains a learned codebook of discrete “patterns”. Each segment of your time series is matched to the nearest codebook entry, converting continuous values into discrete tokens. This allows LLM-style methods to work on time series.

Key Advantages:

  • Converts continuous time series to discrete tokens for LLM compatibility
  • Multiple codebooks capture different aspects of temporal patterns
  • The forecasting stage consumes a tokenizer/codebook trained separately

How It Works

TOTEM learns discrete tokenized representations for time series via VQ-VAE, enabling the use of discrete token-based methods (like LLMs) on continuous time series data.

Reference: Talukder et al., “TOTEM: TOkenized Time Series EMbeddings for General Time Series Analysis”, 2024.

Constructors

Constructor Summary
TOTEMOptions() Initializes a new instance with default values.
TOTEMOptions(TOTEMOptions<T> other) Initializes a new instance by copying from another instance.

Properties

Property Summary
double DropoutRate { get; set; } Gets or sets the dropout rate for regularization.
double LearningRate { get; set; } Gets or sets the peak Adam learning rate for forecasting.
int CodebookDimension { get; set; } Gets or sets the dimension of each codebook entry.
int CodebookSize { get; set; } Gets or sets the number of entries in each codebook.
int ContextLength { get; set; } Gets or sets the number of historical time steps used as input context.
int ForecastHorizon { get; set; } Gets or sets the number of future time steps to forecast.
int HiddenDimension { get; set; } Gets or sets the hidden dimension of the transformer layers.
int NumCodebooks { get; set; } Gets or sets the number of parallel codebooks (product quantization).
int NumHeads { get; set; } Gets or sets the number of attention heads.
int NumLayers { get; set; } Gets or sets the number of transformer layers.
int TotalTrainingSteps { get; set; } Gets or sets the number of optimizer steps in the forecasting OneCycle schedule.