RustMLP

From scratch implementation in Rust to train and run inference in a Multilayer Perceptron (MLP).

Features

• Statistical and Linear algebra functions
• Dataset manipulation
• MLP neural network architecture
• Training and inference loop
• Save/load trained models

Only two dependences:

• rand crate for random number generation
• serde + serde_json crates to serialize/deserialize trained models

Credits

Author: Felipe Gutiérrez Carilao

Example dataset available in Kaggle, cleaned version by me.