Struct TransformerDecoderLayer 

pub struct TransformerDecoderLayer { /* private fields */ }

Transformer Decoder Layer.

Like encoder but with an additional cross-attention layer.

Implementations

impl TransformerDecoderLayer

pub fn new( d_model: usize, nhead: usize, dim_feedforward: usize, ) -> TransformerDecoderLayer

Create a new Transformer Decoder Layer.

pub fn forward_with_memory( &self, tgt: &Tensor, memory: &Tensor, tgt_mask: Option<&Tensor>, memory_mask: Option<&Tensor>, ) -> Tensor

Forward with memory from encoder.

Arguments

• tgt - Target sequence [batch, tgt_len, d_model]
• memory - Encoder output [batch, src_len, d_model]
• tgt_mask - Optional causal mask for target
• memory_mask - Optional mask for encoder-decoder attention

Trait Implementations

impl Debug for TransformerDecoderLayer

fn fmt(&self, f: &mut Formatter<'_>) -> Result<(), Error>

Formats the value using the given formatter.

impl Module for TransformerDecoderLayer

fn forward(&self, input: &Tensor) -> Tensor

Perform forward computation.

fn parameters(&self) -> Vec<&Tensor>

Get references to all learnable parameters.

fn parameters_mut(&mut self) -> Vec<&mut Tensor>

Get mutable references to all learnable parameters.

fn train(&mut self)

Set the module to training mode.

fn eval(&mut self)

Set the module to evaluation mode.

fn training(&self) -> bool

Check if the module is in training mode.

fn named_parameters(&self) -> Vec<(String, &Tensor)>

Get (name, parameter) pairs for all learnable parameters.

fn named_parameters_mut(&mut self) -> Vec<(String, &mut Tensor)>

Get (name, mutable parameter) pairs for all learnable parameters.

fn set_training(&mut self, training: bool)

Set training mode recursively, propagating into child modules.

fn refresh_caches(&mut self)

Refresh any cached computations after parameters have been modified.

fn zero_grad(&mut self)

Zero out gradients for all parameters.

fn num_parameters(&self) -> usize

Get the number of learnable parameters.