v1.0

.gitignore

@@ -0,0 +1 @@
+/target

Cargo.toml

@@ -0,0 +1,9 @@
+cargo-features = ["edition2024"]
+
+[package]
+name = "RustML"
+version = "1.0.0"
+edition = "2024"
+authors = ["Mitchell Shibilski-Unkel"]
+
+[dependencies]

README.md

@@ -1,2 +1,26 @@
 # RustML
 Open-source AI Library for Rust
+
+## About
+`RustML` is an AI/machine learning library built for the `Rust` programming language. It's desgined to work similarly to `Python` libraries, such as PyTorch & Tensorflow. 
+
+## Current Functions & Features
+- Sigmod function
+```rust
+    Sigmod(x: i32) -> i32
+```
+
+- ReLU function
+```rust
+    ReLU(x: i32) -> i32
+```
+
+- Softmax function
+```rust
+    Softmax(x: &[i32]) -> f64
+```
+
+- RNN algoritm
+```rust
+    RNN(x: &[i32], y: &[i32], weights: &[i32], bias: i32) -> Vec<i32> 
+```

src/example.rs

@@ -0,0 +1,9 @@
+mod rustml;
+
+fn main() {
+    let RML = rustml::RustML;
+    let test: [i32; 4] = [1, 2, 3, 4];
+    let testy: [i32; 4] = [5, 6, 7, 8];
+    let testw: [i32; 4] = [55, 66, 77, 88];
+    println!("{:?}", RML.RNN(&test, &testy, &testw, 1));
+}

src/rustml.rs

@@ -0,0 +1,39 @@
+pub struct RustML;
+impl RustML {
+    fn sum(x: &[i32]) -> i32 {
+        x.iter().sum()
+    }
+
+    fn exp(x: i32) -> i32 {
+        f64::exp(x as f64) as i32
+    }
+
+    pub fn Sigmod(x: i32) -> i32 {
+        1 / (1 + Self::exp(-x)) 
+    }
+
+    pub fn Softmax(x: &[i32]) -> f64 {
+        // Apply exp to each element and then sum the results
+        let exp_x: Vec<i32> = x.iter().map(|&v| Self::exp(v)).collect();
+        let sum_exp_x = Self::sum(&exp_x);
+
+        // Return the Softmax result as a float (f64) to avoid integer division
+        exp_x.iter().map(|&v| v as f64 / sum_exp_x as f64).sum()
+    }
+
+    pub fn ReLU(x: i32) -> i32 {
+        std::cmp::max(0, x)
+    }
+
+    pub fn RNN(&self, x: &[i32], y: &[i32], weights: &[i32], bias: i32) -> Vec<i32> {
+        let mut yt: Vec<i32> = Vec::new();
+
+        // ht (hidden state) calculation using Sigmoid
+        for i in 0..x.len() {
+            let temp_value: i32 = (x[i] * weights[i]) + y[i] + bias;
+            yt.push(Self::Sigmod(temp_value)); 
+        }
+
+        yt
+    }
+}