Skip to content

Latest commit

 

History

History
63 lines (43 loc) · 2.67 KB

README.md

File metadata and controls

63 lines (43 loc) · 2.67 KB
Raw

MLTypes

Overview

MLTypes is a Swift package offering advanced types and extensions for machine learning applications. Focused on providing robust, Swift-native implementations for vectors and matrices, it includes extensions for standard types, enhancing their functionality in ML scenarios.

Features

  • MLVector: Generic vector implementation with customizable element types. Supports initialization, element manipulation, and ML-specific operations.
  • MLMatrix: Generic matrix structure to represent and manipulate 2D data arrays in ML models.
  • Randomizable Protocol: Allows for the random initialization of numeric types, vital for stochastic aspects of ML algorithms.
  • IntegerExtensions & FloatingPointExtensions: Extend native Swift types to fit into ML contexts seamlessly.

Roadmap

  • Performance Optimization: Refine current implementations for better performance, especially for large-scale data.
  • Advanced Operations: Introduce more complex vector and matrix operations, like dot product, cross product, matrix inversion, etc.
  • ML Algorithms Integration: Facilitate integration with common ML algorithms, possibly collaborating with other ML frameworks.
  • Documentation & Examples: Expand documentation with more examples and use cases.
  • Testing & Validation: Develop a comprehensive test suite to ensure reliability and accuracy.

Installation

[Provide installation instructions here]

Usage

Initializing Matrices and Vectors

let matrix1: MLMatrix<Int> = MLMatrix(rows: 2, columns: 3, defaultValue: 5) // Type definition is optional
let matrix2 = MLMatrix([[1, 2, 3], [4, 5, 6]])

let vector1: MLVector<Int> = MLVector([1, 2, 3])
let vector2 = MLVector(repeating: 4, size: 3)

Matrix Operations

let transposedMatrix = matrix1.transpose() // Transpose a matrix
let matrixSum = matrix1 + matrix2 // Add two matrices
let productVector: MLVector<Int> = matrix1 * vector1 // Multiply a matrix by a vector (type definition not required)

Vector Operations

let vectorSum = vector1 + vector2 // Vector addition
let dotProduct = vector1.dot(vector2) // Dot product of two vectors
let scaledVector = vector1.scale(by: 2) // Vector scaling

Advanced Operations

let rowDotProduct = matrix1.dot(with: vector1, atRow: 0) // Dot product of a matrix row and a vector
let vectorMatrixProduct = vector1 * matrix2 // Multiplication of a vector by a matrix

Contributing

Contributions are welcome, especially in areas like performance optimization, expanding mathematical operations, and improving integration with ML algorithms. Please follow the contributing guidelines.