nn-rs 🤖

A simple neural network library written in Rust, focusing on performant matrix operations. This project serves as an exploration of Rust's capabilities for numerical computing, including parallel processing with Rayon.

---

Prerequisites

Ensure you have the Rust toolchain installed on your system. You can install it from the official website: rustup.rs.

---

Building and Running

You can build and run the project using standard Cargo commands. It's recommended to use the --release flag for performance-critical tasks.

To build the project:

cargo build --release

To run the main application binary:

cargo run --release

---

Running Tests 🧪

This project contains both unit tests (located within the src/ directory) and integration tests (located in the tests/ directory).

To run all tests, execute the following command:

cargo test

---

Running Benchmarks ⏱

The project includes a custom benchmark suite to measure and compare the performance of different matrix multiplication algorithms (sequential, parallel, and optimized).

To run the benchmarks, use this command:

cargo bench

---

Project Structure

• src/: Contains the main application (main.rs) and all the core library logic (e.g., matrix.rs).
• tests/: Contains integration tests that test the public API of the library.
• benches/: Contains custom benchmark programs for performance measurement.
• Cargo.toml: The package manifest that defines project metadata and dependencies.