Getting Started & FAQ

Welcome to Born! This pinned discussion answers common questions.

Installation

go get github.com/born-ml/[email protected]

Requirements: Go 1.25+

FAQ

What is Born?

Born is a production-ready ML framework for Go. Train models anywhere, deploy as single binaries.

Why Go instead of Python?

• Single binary deployment (no runtime)
• Fast startup (<100ms)
• Type safety at compile time
• Native integration with Go backends
• Cross-compilation out of the box

Does Born support GPU?

Yes! v0.2.0 includes WebGPU backend with 123x speedup for matrix operations.

Currently Windows only. Linux/macOS GPU support planned for v0.4.0.

Can I use PyTorch models?

Yes, via ONNX import (planned for v0.3.0-v0.4.0). Train in PyTorch, deploy with Born.

How does Born compare to Gorgonia?

Born focuses on:

• Modern Go (generics)
• Zero CGO (including GPU!)
• Production-first design
• Burn-inspired architecture

Where are the examples?

• examples/mnist/ - MLP classifier (97.44% accuracy)
• examples/mnist-cnn/ - CNN classifier (98.18% accuracy)
• examples/mnist-gpu/ - GPU benchmark

Getting Help

1. Search existing discussions first
2. Check examples in the repo
3. Read the docs: pkg.go.dev
4. Ask here with code examples

Links

• README
• CONTRIBUTING
• CHANGELOG
• ROADMAP