SwiftLearn

A native Swift inference library for scikit-learn models. Train your models in Python, run predictions natively in Swift.

Why SwiftLearn?

• Train in Python, Deploy in Swift: Use familiar sklearn workflows, then run inference on Apple platforms without Python dependencies
• Native Performance: Built on Apple's Accelerate framework (vDSP) for vectorized operations
• No Runtime Dependencies: Pure Swift - no Python bridge, no CoreML conversion
• Type Safe: Leverages Swift's type system for compile-time safety

How It Works

1. Train your model in Python using scikit-learn
2. Export model weights to JSON/H5
3. Load weights in Swift and run predict/predict_proba

# Python: Train and export
from sklearn.preprocessing import StandardScaler
import json

scaler = StandardScaler()
scaler.fit(X_train)

# Export weights
weights = {
    "mean_": scaler.mean_.tolist(),
    "scale_": scaler.scale_.tolist()
}
json.dump(weights, open("scaler.json", "w"))

// Swift: Load and predict
import SwiftLearn

let scaler = try StandardScaler.load(from: "scaler.json")
let predictions = try scaler.transform(newData)

Installation

Swift Package Manager

Add to your Package.swift:

dependencies: [
    .package(url: "https://github.com/aaryanshsahay/SwiftLearn.git", branch: "dev")
]

Or in Xcode: File > Add Package Dependencies > Enter the repository URL.

Quick Start

import SwiftLearn

// Create data
let X = Matrix([
    [0.0, 0.0],
    [0.0, 0.0],
    [1.0, 1.0],
    [1.0, 1.0]
])

// StandardScaler example
var scaler = StandardScaler()
let scaled = try scaler.fitTransform(X)

print(scaler.mean_!)   // [0.5, 0.5]
print(scaler.scale_!)  // [0.5, 0.5]

// Transform new data
let newData = Matrix([[2.0, 2.0]])
let result = try scaler.transform(newData)

Available Components

Preprocessing

Class	Description
StandardScaler	Standardize features by removing mean and scaling to unit variance

Core Types

Type	Description
Matrix	2D array with row-major storage, supports subscripting and column access

Roadmap

Preprocessing

• StandardScaler
• MinMaxScaler
• LabelEncoder
• OneHotEncoder

Model Loading

• Load model weights from JSON
• Load model weights from H5

Inference

• Linear Regression (predict)
• Logistic Regression (predict, predict_proba)
• K-Nearest Neighbors (predict, predict_proba)
• Decision Trees (predict, predict_proba)
• Random Forest (predict, predict_proba)
• SVM (predict, decision_function)

Utilities

• Train/Test Split
• Accuracy, Precision, Recall metrics

Requirements

• Swift 5.9+
• macOS 10.15+ / iOS 13+ / tvOS 13+ / watchOS 6+

Contributing

Contributions are welcome! Please open an issue or submit a pull request.

License

MIT License - see LICENSE file for details.