sklears-manifold

Latest release: 0.2.0 (July 14, 2026). See the workspace release notes for highlights and upgrade guidance.

Overview

sklears-manifold implements manifold learning, nonlinear dimensionality reduction, and embedding algorithms mirroring scikit-learn’s manifold module.

Key Features

• Algorithms: t-SNE, UMAP, Isomap, Locally Linear Embedding, Spectral Embedding, MDS.
• Performance: Barnes-Hut t-SNE, GPU-accelerated pairwise distances, HNSW-backed approximate kNN search for Euclidean/Cosine metrics (gpu feature, via oxicuda-manifold), and parallel exact kNN search (rayon).
• Visualization: Embedding utilities that integrate with sklears-inspection and Python plotting stacks.
• Pipeline Support: Works seamlessly with preprocessing, decomposition, and clustering crates.

Quick Start

use sklears_manifold::TSNE;
use sklears_core::traits::{Fit, Transform};
use scirs2_core::ndarray::array;

let x = array![
    [0.0, 0.0], [1.0, 1.0], [2.0, 2.0],
    [10.0, 10.0], [11.0, 11.0], [12.0, 12.0],
];

let tsne = TSNE::new()
    .n_components(2)
    .perplexity(2.0)
    .n_iter(100);

let fitted = tsne.fit(&x.view(), &())?;
let embedding = fitted.embedding();

Status

• Validated by 422 passing crate tests for 0.2.0 (2 skipped).
• Performance parity (and in many cases superiority) compared with scikit-learn’s manifold implementations.
• Upcoming tasks (GPU UMAP, streaming embeddings) tracked in TODO.md.