rlx-umap 0.2.5

Parametric UMAP on RLX — full fit/transform API and k-NN building blocks.
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
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144

// RLX — versatile ML compiler + runtime.
// Copyright (C) 2026 Eugene Hauptmann, Nataliya Kosmyna.
//
// This program is free software: you can redistribute it and/or modify
// it under the terms of the GNU General Public License as published by
// the Free Software Foundation, version 3.
//
// This program is distributed in the hope that it will be useful,
// but WITHOUT ANY WARRANTY; without even the implied warranty of
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
// GNU General Public License for more details.
//
// You should have received a copy of the GNU General Public License
// along with this program. If not, see <https://www.gnu.org/licenses/>.

//! Parametric UMAP on RLX — k-NN building blocks + full [`Umap::fit`](umap::Umap::fit) API matching [fast-umap](https://github.com/eugenehp/fast-umap).
//!
//! - [`Umap::fit`] / [`fit_with_progress`](umap::Umap::fit_with_progress) — sparse cross-entropy training
//! - [`FittedUmap::transform`] — inference with training z-score stats
//! - [`FittedUmap::save`] / [`load`] — safetensors or GGUF (`.ruama` load only for legacy files)
//!
//! Call [`register`] once per process before Session execution.
//!
//! ## Quick start
//!
//! ```ignore
//! rlx_umap::register();
//!
//! use rlx_ir::{DType, Graph, Shape};
//! use rlx_umap::knn_indices_and_distances;
//!
//! let mut g = Graph::new("knn");
//! let pairwise = g.input("pairwise", Shape::new(&[64, 64], DType::F32));
//! let (idx, dist) = knn_indices_and_distances(&mut g, pairwise, 5);
//! g.set_outputs(vec![idx, dist]);
//! ```

pub mod knn;
pub mod knn_attrs;
pub mod pack;
pub mod pairwise;
pub mod parity;

#[cfg(feature = "cpu")]
pub mod ops;

#[cfg(feature = "graph")]
pub mod graph;

#[cfg(feature = "bench")]
pub mod session;

#[cfg(feature = "full")]
pub mod adam;
#[cfg(feature = "full")]
pub mod config;
#[cfg(feature = "full")]
pub mod data;
#[cfg(feature = "full")]
pub mod encoder;
#[cfg(feature = "full")]
pub mod fitted;
#[cfg(feature = "full")]
pub mod interrupt;
#[cfg(feature = "full")]
pub mod model;
#[cfg(feature = "full")]
pub mod model_io;
#[cfg(feature = "nn-descent")]
pub mod nn_descent;
#[cfg(feature = "pca")]
pub mod pca;
#[cfg(feature = "full")]
pub mod prelude;
#[cfg(feature = "full")]
pub mod serialize;
#[cfg(feature = "full")]
pub mod train;
#[cfg(feature = "full")]
pub mod training;
#[cfg(feature = "full")]
pub mod umap;
#[cfg(feature = "full")]
pub mod utils;
#[cfg(feature = "full")]
pub mod weights;

#[cfg(all(feature = "optim", feature = "full"))]
pub mod optim_adapter;

#[cfg(all(feature = "metal", target_os = "macos"))]
mod metal_kernels;

#[cfg(all(feature = "mlx", target_os = "macos"))]
mod mlx_kernels;

pub use knn::{knn_backward_pairwise, knn_forward_packed};
pub use knn_attrs::KnnAttrs;
pub use pack::unpack_knn_packed;
pub use pairwise::{cosine_pairwise_reference, euclidean_pairwise_reference};
pub use parity::{KnnParityReport, compare_knn, max_pairwise_error};

#[cfg(feature = "cpu")]
pub use ops::{UMAP_KNN, UMAP_KNN_BWD, register_umap_ops};

#[cfg(feature = "graph")]
pub use graph::{
    cosine_knn_graph, cosine_knn_packed_graph, knn_graph, knn_indices_and_distances,
    pairwise_cosine_graph, pairwise_euclidean_graph, split_knn_packed,
};

/// Register UMAP custom ops (IR + CPU). Alias of [`register_umap_ops`].
#[cfg(feature = "cpu")]
pub fn register() {
    register_umap_ops();
}

#[cfg(feature = "full")]
pub use config::UmapConfig;
#[cfg(feature = "full")]
pub use data::{load_csv, load_f64_matrix, load_synthetic, write_embedding_csv};
#[cfg(feature = "full")]
pub use fitted::FittedUmap;
#[cfg(feature = "full")]
pub use model_io::{
    EXT_GGUF, EXT_RUAMA, EXT_SAFETENSORS, MODEL_EXT, format_from_path, model_path,
    model_path_with_ext, weight_shapes,
};
#[cfg(feature = "full")]
pub use rlx_driver::Device;
#[cfg(feature = "full")]
pub use serialize::{
    LoadedModel, ModelMetadata, SaveBundle, load_model, load_weights, save_model, save_weights,
};
#[cfg(feature = "full")]
pub use train::EpochProgress;
#[cfg(feature = "full")]
pub use training::{FitOptions, TrainResult, fit, fit_with_progress, train_only};
#[cfg(feature = "full")]
pub use umap::Umap;
#[cfg(feature = "full")]
pub use utils::NormStats;
#[cfg(feature = "full")]
pub use weights::WeightStore;