Struct LearnedPositionEncoding 

pub struct LearnedPositionEncoding {
    pub config: PositionEncodingConfig,
}

Learned position encoding

Fields

config: PositionEncodingConfig

Configuration

Implementations

impl LearnedPositionEncoding

pub fn new(config: PositionEncodingConfig) -> Result<Self>

Create a new learned position encoding

pub fn build_encoding_graph( &self, graph: &mut EinsumGraph, ) -> Result<Vec<usize>>

Build einsum graph for learned position encoding

Input tensors:

• 0: x (input) [batch, seq_len, d_model]
• 1: position_embeddings [max_seq_len, d_model] (learned parameter)
• 2: position_ids [batch, seq_len] (optional)

Output tensors:

• output: [batch, seq_len, d_model] (x + position_embedding)

pub fn max_seq_len(&self) -> usize

Get maximum sequence length

Trait Implementations

impl Clone for LearnedPositionEncoding

fn clone(&self) -> LearnedPositionEncoding

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for LearnedPositionEncoding

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

Formats the value using the given formatter.