Tensor Frame

A high-performance, PyTorch-like tensor library for Rust with support for multiple computational backends.

Features

🚀 Multiple Backends: CPU (Rayon), WGPU, and CUDA support
🔄 Automatic Backend Selection: Falls back to best available backend
📐 Broadcasting: NumPy/PyTorch-style automatic broadcasting
🎯 Type Safety: Rust's type system for memory safety
⚡ Zero-Copy Operations: Efficient memory management
🎛️ Feature Flags: Optional dependencies for different backends

Quick Start

Add to your Cargo.toml:

[dependencies]
tensor_frame = "0.0.1-alpha"

# For GPU support
tensor_frame = { version = "0.0.1-alpha", features = ["wgpu"] }

Basic usage:

use tensor_frame::Tensor;

// Create tensors (automatically uses best backend)
let a = Tensor::ones(vec![2, 3])?;
let b = Tensor::zeros(vec![2, 3])?;

// Operations with broadcasting
let c = (a + b)?;
let sum = c.sum(None)?;

println!("Result: {:?}", sum.to_vec()?);

Backends

CPU Backend (Default)

Uses Rayon for parallel computation
Always available
Good for small to medium tensors

WGPU Backend

Cross-platform GPU compute
Supports Metal, Vulkan, DX12, OpenGL
Enable with features = ["wgpu"]

CUDA Backend

NVIDIA GPU acceleration
Enable with features = ["cuda"]
Requires CUDA toolkit

Documentation

📖 Complete Guide - Comprehensive documentation with tutorials
🚀 Getting Started - Quick start guide
📚 API Reference - Detailed API documentation
💡 Examples - Practical examples and tutorials
⚡ Performance Guide - Optimization tips and benchmarks
🔧 Backend Guides - CPU, WGPU, and CUDA backend details

Examples

See the examples directory for more detailed usage:

Contributing

Contributions are welcome! Please see CONTRIBUTING.md for guidelines.

License

Licensed under either of

Apache License, Version 2.0 (LICENSE-APACHE)
MIT License (LICENSE-MIT)

at your option.

tensor_frame 0.0.1-alpha