optirs-learned 0.3.2

OptiRS learned optimizers and meta-learning
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19

// Transformer architecture components
//
// This module contains the core architectural components of the transformer optimizer,
// organized into separate submodules for better maintainability and testing.

pub mod attention;
pub mod encoder;
pub mod feedforward;
pub mod positional_encoding;

// Re-export key types for convenience
pub use attention::{
    AttentionOptimization, MultiHeadAttention, RelativePositionBias, RoPEEmbeddings,
};
pub use encoder::{
    ActivationFunction, DropoutLayer, FeedForwardNetwork, LayerNorm, TransformerLayer,
};
pub use feedforward::{InputEmbedding, OutputProjectionLayer, OutputTransformation};
pub use positional_encoding::{PositionalEncoder, PositionalEncodingType};