XLSTMOptions

Options & Configuration · AiDotNet.NeuralNetworks.Options

Configures the dimensions and default optimizer learning rate of the xLSTM language model.

Inheritance: ModelOptions → NeuralNetworkOptions → ModelHyperparameterOptions → SequenceModelOptions → XLSTMOptions

For Beginners

xLSTM extends the familiar recurrent memory cell with more expressive gates and memory. Width, depth, and head count control the model’s capacity; the learning rate controls the size of each training step.

How It Works

Beck et al., xLSTM: Extended Long Short-Term Memory (2024), introduce exponential gating, scalar-memory sLSTM, and matrix-memory mLSTM blocks. The original language-model experiments include 125M, 350M, 760M, and 1.3B configurations. These options retain the library’s 256-wide, four-layer defaults rather than claiming checkpoint equivalence, and do not configure every block-mixture choice in the paper.

Constructors

Constructor Summary
XLSTMOptions() Initializes a new instance of the XLSTMOptions class carrying this model’s shipped defaults.
XLSTMOptions(XLSTMOptions other) Initializes an instance by copying every declared and inherited setting.

Properties

Property Summary
double LearningRate { get; set; } AdamW learning rate.