TensorLab

A learning-focused Rust tensor library for exploring linear algebra operations. Built from scratch as a hands-on approach to understanding matrix mathematics and numerical computing.

Project Goal

This library serves as a practical companion to learning linear algebra. Each operation is implemented manually to understand the underlying mechanics of tensor computations.

Current Features

Core Operations

• Element-wise operations: Addition, subtraction, Hadamard product (element-wise multiplication)
• Scalar multiplication: Multiply tensors by scalar values
• Matrix multiplication: Standard matrix product (matmul)
• Operator overloading: Use +, -, * operators naturally

Data Management

• Row-major tensor storage with shape validation
• Element access via coordinates
• Comprehensive error handling

Quick Start

use tensor_zero::Tensor;

// Create tensors
let a = Tensor::new(&vec![1.0, 2.0, 3.0, 4.0], &vec![2, 2]).unwrap();
let b = Tensor::new(&vec![5.0, 6.0, 7.0, 8.0], &vec![2, 2]).unwrap();

// Element-wise operations
let sum = (&a + &b).unwrap();     // [6.0, 8.0, 10.0, 12.0]
let diff = (&a - &b).unwrap();    // [-4.0, -4.0, -4.0, -4.0]
let prod = (&a * &b).unwrap();    // [5.0, 12.0, 21.0, 32.0]

// Scalar multiplication
let scaled = a.mul_scalar(2.0);   // [2.0, 4.0, 6.0, 8.0]

// Matrix multiplication
let x = Tensor::new(&vec![1.0, 2.0, 3.0, 4.0, 5.0, 6.0], &vec![2, 3]).unwrap();
let y = Tensor::new(&vec![7.0, 8.0, 9.0, 1.0, 2.0, 3.0], &vec![3, 2]).unwrap();
let result = x.matmul(&y).unwrap(); // Shape: [2, 2]

🏗️ Project Structure

tensorlab/
tensorlab/
├── src/
│   ├── lib.rs                    # Public API and module declarations
│   ├── error.rs                  # Error types and Display implementation
│   ├── tensor.rs                 # Tensor struct and core operations
│   ├── operations.rs             # Operator trait implementations
│   ├── lib.rs                    # Public API and module declarations
│   ├── error.rs                  # Error types and Display implementation
│   ├── tensor.rs                 # Tensor struct and core operations
│   ├── operations.rs             # Operator trait implementations
│   └── main.rs                   # Example usage
├── tests/
│   └── integration_test.rs       # Comprehensive test suite
├── Cargo.toml                    # Project manifest
└── README.md                     # This file

Module Breakdown

• error.rs: Defines the Error enum for shape mismatches and invalid operations
• tensor.rs: Core Tensor struct with methods for:
  • Construction and element access
  • Element-wise operations (add, sub, mul)
  • Scalar multiplication
  • Matrix multiplication
• operations.rs: Implements Add, Sub, and Mul traits for ergonomic operator usage
• lib.rs: Consolidates and re-exports public API

Module Breakdown

• error.rs: Defines the Error enum for shape mismatches and invalid operations
• tensor.rs: Core Tensor struct with methods for:
  • Construction and element access
  • Element-wise operations (add, sub, mul)
  • Scalar multiplication
  • Matrix multiplication
• operations.rs: Implements Add, Sub, and Mul traits for ergonomic operator usage
• lib.rs: Consolidates and re-exports public API

Testing

cargo test

All operations include integration tests covering both valid operations and error cases.

Roadmap

Next Steps:

• Matrix transpose
• Optimize matmul with transpose
• Parallel operations with Rayon

Future Goals:

• Broadcasting support
• N-dimensional operations
• Reshape and view operations
• Advanced linear algebra (inverse, determinant, SVD)

Technical Details

• Storage: Flat Vec<f32> with row-major ordering
• Operations: Iterator-based functional style
• Errors: Typed errors for shape mismatches and invalid operations
• Memory: Zero-copy references where possible

🤝 Contributing

This is primarily a learning project, but feel free to suggest improvements or share your own learning journey!

Note: Active development follows my linear algebra learning progress. Expect frequent updates as new concepts are explored.