ruda-nn 0.21.23

Ruda neural network layers, activation modules and losses.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47

/// Attention module
pub mod attention;

/// Cache module
pub mod cache;

/// Convolution module
pub mod conv;

/// Pooling module
pub mod pool;

/// Transformer module
pub mod transformer;

/// Interpolate module
pub mod interpolate;

mod dropout;
mod embedding;
mod linear;
mod lora;
/// Manifold-constrained hyper-connections and residual mixing.
pub mod mhc;
#[cfg(feature = "sparse")]
mod sparse_linear;
mod noise;
mod pos_encoding;
mod rnn;
mod rope_encoding;
mod unfold;

pub mod norm;
pub use norm::{batch::*, group::*, instance::*, layer::*, local_response::*, rms::*};

pub use dropout::*;
pub use embedding::*;
pub use linear::*;
pub use lora::*;
pub use mhc::*;
#[cfg(feature = "sparse")]
pub use sparse_linear::*;
pub use noise::*;
pub use pos_encoding::*;
pub use rnn::*;
pub use rope_encoding::*;
pub use unfold::*;