Trait tch::nn::RNN

pub trait RNN {
    type State;
    fn zero_state(&self, batch_dim: i64) -> Self::State;
    fn step(&self, input: &Tensor, state: &Self::State) -> Self::State;
    fn seq_init(
        &self, 
        input: &Tensor, 
        state: &Self::State
    ) -> (Tensor, Self::State);

    fn seq(&self, input: &Tensor) -> (Tensor, Self::State) { ... }
}

Trait for Recurrent Neural Networks.

Associated Types

type State

Loading content...

Required methods

fn zero_state(&self, batch_dim: i64) -> Self::State

A zero state from which the recurrent network is usually initialized.

fn step(&self, input: &Tensor, state: &Self::State) -> Self::State

Applies a single step of the recurrent network.

The input should have dimensions [batch_size, features].

fn seq_init(&self, input: &Tensor, state: &Self::State) -> (Tensor, Self::State)

Applies multiple steps of the recurrent network.

The input should have dimensions [batch_size, seq_len, features].

Loading content...

Provided methods

fn seq(&self, input: &Tensor) -> (Tensor, Self::State)

Applies multiple steps of the recurrent network.

The input should have dimensions [batch_size, seq_len, features]. The initial state is the result of applying zero_state.

Loading content...

Implementors

impl RNN for GRU

type State = GRUState

fn zero_state(&self, batch_dim: i64) -> GRUState

fn step(&self, input: &Tensor, in_state: &GRUState) -> GRUState

fn seq_init(&self, input: &Tensor, in_state: &GRUState) -> (Tensor, GRUState)

impl RNN for LSTM

type State = LSTMState

fn zero_state(&self, batch_dim: i64) -> LSTMState

fn step(&self, input: &Tensor, in_state: &LSTMState) -> LSTMState

fn seq_init(&self, input: &Tensor, in_state: &LSTMState) -> (Tensor, LSTMState)

Loading content...