Lumina - Rust Tensor Framework

Lumina combines Rust's memory safety with GPU acceleration through vulkan and CUDA backends. The framework supports both f32 and f16 precision modes, enabling efficient operations for both training and inference use cases.

Performance Benchmarks

Stress testing matrix operations across different device configurations:

Core Features

• • Safe, type-checked tensor operations with compile-time shape verification
• • 2.4x speed improvement over numpy for large matrix calculations
• • Automatic GPU memory pooling across vulkan/CUDA devices
• • Python interoperability via PyO3 bindings

Code Example

Basic tensor operations in Lumina:

#[derive(Debug)]
struct Tensor {
    data: Vec,
    shape: Vec,
}

impl + Copy> Tensor {
    fn add(&self, other: &Self) -> Self {
        assert_eq!(self.shape, other.shape);
        
        Tensor {
            data: self.data.iter()
                        .zip(other.data.iter())
                        .map(|(a, b)| *a + *b)
                        .collect(),
            shape: self.shape.clone(),
        }
    }
}