smartcore
A fast, ergonomic machine learning library for Rust with broad algorithm coverage and WASM-first defaults.
What is smartcore?
smartcore is a comprehensive machine learning library for Rust that provides production-friendly APIs, strong typing, and good defaults while remaining flexible for research and experimentation. It covers classical supervised and unsupervised methods with a modular linear algebra abstraction and optional ndarray support.
Rust developers building machine learning applications for production, especially those targeting WebAssembly or embedded deployments, as well as researchers needing a flexible, idiomatic Rust ML library.
Developers choose smartcore for its WASM-first design tuned for portability, ergonomic API with consistent traits, and broad algorithm coverage including linear models, tree-based methods, ensembles, SVMs, and clustering, all while maintaining strong typing and modular linear algebra.
Overview
A comprehensive library for machine learning and numerical computing. Apply Machine Learning with Rust leveraging first principles.
Use Cases
Best For
- Building machine learning applications in Rust that need to run in WebAssembly or embedded environments.
- Implementing classical supervised learning algorithms like linear regression, logistic regression, and SVM in a type-safe Rust API.
- Implementing classical unsupervised learning algorithms like K-Means, DBSCAN, and hierarchical clustering in Rust.
- Developing ML pipelines with practical utilities for model selection, evaluation, and data preprocessing in Rust.
- Experimenting with ML algorithms in Jupyter notebooks using Rust via the companion smartcore-jupyter repository.
- Projects requiring a modular linear algebra abstraction with optional ndarray integration for array-first workflows.
Not Ideal For
- Teams primarily using Python who need extensive deep learning or neural network frameworks like TensorFlow or PyTorch.
- Projects requiring out-of-the-box GPU acceleration for training or inference, as smartcore's focus is on CPU-based classical methods.
- Applications dependent on a large ecosystem of pre-trained models or third-party integrations common in mature ML libraries like scikit-learn.
- Organizations needing stable, long-term APIs without breaking changes, given smartcore's recent refactors and deprecations in version 0.4.
Pros & Cons
Pros
Defaults are tuned for WebAssembly and embedded deployments, making it ideal for browser-based or resource-constrained environments, with features like serde and datasets opt-in to minimize footprint.
Covers classical supervised and unsupervised methods, including linear models, tree-based ensembles, SVMs, and clustering, as highlighted in the Algorithms section for comprehensive ML workflows.
Provides consistent traits across modules and examples using native Rust vectors and matrices, with a quick start example demonstrating ease of use for tasks like KNN classification.
Offers strong linear algebra traits with optional ndarray integration, allowing flexibility for array-first workflows while maintaining abstraction, as noted in the Highlights.
Cons
Focuses solely on classical machine learning algorithms, lacking neural networks or deep learning capabilities, which limits its use for cutting-edge AI tasks.
Recent version updates like v0.4 introduced breaking changes and deprecations, such as dropping nalgebra-bindings, requiring migration efforts and potentially disrupting existing codebases.
As a Rust library, it has a nascent community compared to Python-based alternatives, which can mean fewer third-party tools, pre-trained models, or community support resources.
Frequently Asked Questions
Related Projects
PyTorch - Tensors and Dynamic neural networks in Python with strong GPU accelerationTensors and Dynamic neural networks in Python with strong GPU acceleration
kerasDeep Learning for humans
streamlitStreamlit — A faster way to build and share data apps.
gradioBuild and share delightful machine learning apps, all in Python. 🌟 Star to support our work!
Found a gem we're missing?
Open-Awesome is built by the community, for the community. Submit a project, suggest an awesome list, or help improve the catalog on GitHub.