zyx-nn 0.15.3

Zyx nn modules
Documentation 
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
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77

// Copyright (C) 2025 zk4x
// SPDX-License-Identifier: LGPL-3.0-only

//! nn modules for zyx ML library

#![forbid(unsafe_code)]
#![forbid(rustdoc::broken_intra_doc_links)]
#![forbid(rustdoc::private_intra_doc_links)]
#![forbid(missing_docs)]
#![forbid(rustdoc::missing_crate_level_docs)]
//#![forbid(rustdoc::missing_doc_code_examples)]
#![forbid(rustdoc::private_doc_tests)]
#![forbid(rustdoc::invalid_codeblock_attributes)]
#![forbid(rustdoc::invalid_html_tags)]
#![forbid(rustdoc::invalid_rust_codeblocks)]
#![forbid(rustdoc::bare_urls)]
#![forbid(rustdoc::unescaped_backticks)]
#![forbid(rustdoc::redundant_explicit_links)]

pub use zyx_derive::Module;

mod linear;
pub use linear::Linear;

// Normalization layers
mod layer_norm;
pub use layer_norm::LayerNorm;

mod batch_norm;
pub use batch_norm::BatchNorm;

mod group_norm;
pub use group_norm::GroupNorm;

mod rms_norm;
pub use rms_norm::RMSNorm;

// Recurrent layers
mod rnn;
pub use rnn::RNNCell;

mod gru;
pub use gru::GRUCell;

mod lstm;
pub use lstm::LSTMCell;

mod causal_self_attention;
pub use causal_self_attention::CausalSelfAttention;

mod multi_head_attention;
pub use multi_head_attention::MultiheadAttention;

mod embedding;
pub use embedding::Embedding;

mod conv2d;
pub use conv2d::Conv2d;

mod positional_encoding;
pub use positional_encoding::PositionalEncoding;

mod transformer_encoder_layer;
pub use transformer_encoder_layer::TransformerEncoderLayer;

mod transformer_decoder_layer;
pub use transformer_decoder_layer::TransformerDecoderLayer;

// I don't like this, use fn(&Tensor) -> Tensor or something like that instead
// fn(impl Into<Tensor>) -> Tensor
//mod activation;
//pub use activation::Activation;

#[cfg(feature = "py")]
mod py_bindings;
#[cfg(feature = "py")]
pub use py_bindings::register_nn;