Struct rogue_net::RogueNetConfig

pub struct RogueNetConfig {
    pub embd_pdrop: f64,
    pub resid_pdrop: f64,
    pub attn_pdrop: f64,
    pub n_layer: u32,
    pub n_head: u32,
    pub d_model: u32,
    pub pooling: Option<String>,
    pub relpos_encoding: Option<RelposEncodingConfig>,
    pub d_qk: u32,
    pub translation: Option<TranslationConfig>,
}

Network architecture hyperparameters for RogueNet.

Fields

embd_pdrop: f64

Dropout probability for the embedding layer.

resid_pdrop: f64

Dropout probability on attention block output.

attn_pdrop: f64

Dropout probability on attention probabilities.

n_layer: u32

Number of transformer blocks.

n_head: u32

Number of attention heads.

d_model: u32

Model width.

pooling: Option<String>

Replace attention with a pooling layer.

relpos_encoding: Option<RelposEncodingConfig>

Settings for relative position encoding.

d_qk: u32

Width of keys and queries used in entity-selection heads.

translation: Option<TranslationConfig>

Configuration for translating positions of all entities with respect to a reference entity.

Trait Implementations

impl Clone for RogueNetConfig

fn clone(&self) -> RogueNetConfig

Returns a copy of the value.

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

Performs copy-assignment from source.

impl Debug for RogueNetConfig

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

Formats the value using the given formatter.

impl<'de> Deserialize<'de> for RogueNetConfig

fn deserialize<__D>(__deserializer: __D) -> Result<Self, __D::Error>where
    __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl Serialize for RogueNetConfig

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

Serialize this value into the given Serde serializer.