AutoregressiveModel

kizzasi_model

Trait AutoregressiveModel 

Source 
pub trait AutoregressiveModel: SignalPredictor + Send {
    // Required methods
    fn hidden_dim(&self) -> usize;
    fn state_dim(&self) -> usize;
    fn num_layers(&self) -> usize;
    fn model_type(&self) -> ModelType;
    fn get_states(&self) -> Vec<HiddenState>;
    fn set_states(&mut self, states: Vec<HiddenState>) -> ModelResult<()>;
}
Expand description

Trait for model architectures that support autoregressive prediction

Required Methods§

Source

fn hidden_dim(&self) -> usize

Get the model’s hidden dimension

Source

fn state_dim(&self) -> usize

Get the model’s state dimension (for SSMs)

Source

fn num_layers(&self) -> usize

Get number of layers

Source

fn model_type(&self) -> ModelType

Get model type identifier

Source

fn get_states(&self) -> Vec<HiddenState>

Get current hidden states for all layers

Source

fn set_states(&mut self, states: Vec<HiddenState>) -> ModelResult<()>

Set hidden states for all layers

Implementors§