argmin-rs/argmin

Health v2 · maintenance only

Adoption not part of the score

1275 stars · 116 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

argmin is a numerical optimization library written entirely in pure Rust, offering a wide range of optimization algorithms behind a consistent, type-agnostic interface. It supports pluggable math backends like ndarray and nalgebra, observers for tracking iteration progress, and optional checkpointing for resilience.

Use cases

• minimize a cost function in Rust
• run gradient descent or Newton-type solvers
• fit a model by numerical optimization
• implement a custom optimization solver with a consistent framework
• swap between ndarray and nalgebra math backends
• resume long optimization runs after crashes via checkpointing

When to choose

• you need optimization algorithms in pure Rust without C/Fortran dependencies
• you want a consistent interface to try multiple solvers on the same problem
• you need type-agnostic math with your own types or backend
• you want observers and checkpointing for long-running optimizations

When to avoid

• you need a batteries-included scientific computing stack like SciPy
• you work primarily in Python or another language
• you need symbolic or mixed-integer optimization rather than numerical continuous optimization

Facets

library · maturity active

math machine-learning mathematics data-science performance rust cross-platform numerical-optimization optimization-algorithms solvers gradient-descent type-agnostic checkpointing algorithms

Member repositories

For agents

markdown · JSON · MCP: product_card(name="argmin-rs/argmin")