Struct Model 

pub struct Model {
    pub vocab_size: usize,
    pub embedding_dim: usize,
    pub input_weights: Vec<f32>,
    pub output_weights: Vec<f32>,
}

Core weight matrices for Word2Vec.

Both matrices are row-major flat Vec<f32> for cache efficiency. Row i starts at byte offset i * embedding_dim * 4.

Fields

vocab_size: usize

embedding_dim: usize

input_weights: Vec<f32>

output_weights: Vec<f32>

Implementations

impl Model

pub fn new(vocab_size: usize, embedding_dim: usize, seed: u64) -> Self

Initialise with Xavier uniform weights.

use word2vec::model::Model;
let m = Model::new(100, 50, 42);
assert_eq!(m.input_weights.len(), 100 * 50);

pub fn input_vec(&self, idx: usize) -> &[f32]

Get the embedding vector for word at idx (slice into input_weights).

pub fn input_vec_mut(&mut self, idx: usize) -> &mut [f32]

Mutable access to input embedding.

pub fn output_vec_mut(&mut self, idx: usize) -> &mut [f32]

Mutable access to output embedding.

pub fn skipgram_update( &mut self, center: usize, context: usize, negatives: &[usize], lr: f32, ) -> f32

Skip-gram update: given center word, update for one context word and n_neg negative samples.

Returns the binary cross-entropy loss contribution.

pub fn cbow_update( &mut self, center: usize, context_words: &[usize], negatives: &[usize], lr: f32, ) -> f32

CBOW update: average context embeddings, predict center word.

pub fn update( &mut self, model_type: ModelType, center: usize, context_window: &[usize], negatives: &[usize], lr: f32, ) -> f32

Run a full Skip-gram or CBOW update, dispatching on model type.

Trait Implementations

impl Clone for Model

fn clone(&self) -> Model

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for Model

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

Formats the value using the given formatter.

impl<'de> Deserialize<'de> for Model

fn deserialize<__D>(__deserializer: __D) -> Result<Self, __D::Error>

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl Serialize for Model

fn serialize<__S>(&self, __serializer: __S) -> Result<__S::Ok, __S::Error>

where __S: Serializer,

Serialize this value into the given Serde serializer.