MarioSieg/magnetron
A machine learning runtime written in C with a Python API, implementing tensors, autograd, operators, and execution engine from scratch.
Not currently ranked — collecting fresh signals.
star history
Magnetron is a compact ML framework that reimplements core ML infrastructure including tensor system, autograd engine, operators, and execution model in C with a Python interface. It targets developers who want full control over execution and memory without the complexity of larger frameworks. Includes example implementations for running real models including Qwen3 and GPT-2 inference.
Frequently asked
- What is MarioSieg/magnetron?
- A machine learning runtime written in C with a Python API, implementing tensors, autograd, operators, and execution engine from scratch.
- Is magnetron open source?
- Yes — MarioSieg/magnetron is an open-source project tracked on heatdrop.
- What language is magnetron written in?
- MarioSieg/magnetron is primarily written in C.
- How popular is magnetron?
- MarioSieg/magnetron has 693 stars on GitHub.
- Where can I find magnetron?
- MarioSieg/magnetron is on GitHub at https://github.com/MarioSieg/magnetron.