luminal-ai/luminal
A Rust-based inference compiler that optimizes and runs ML models on GPU hardware.
Not currently ranked — collecting fresh signals.
star history
Luminal is a general-purpose inference compiler for ML models written in Rust. It directly integrates with PyTorch as a compiler backend via torch.compile(), allowing users to compile PyTorch models for optimized inference on CUDA hardware. The project targets near-theoretical maximum performance on hardware like H100 GPUs and supports running large language models like Llama 3 8B locally on consumer and data-center GPUs.
Frequently asked
- What is luminal-ai/luminal?
- A Rust-based inference compiler that optimizes and runs ML models on GPU hardware.
- Is luminal open source?
- Yes — luminal-ai/luminal is open source, released under the Apache-2.0 license.
- What language is luminal written in?
- luminal-ai/luminal is primarily written in Rust.
- How popular is luminal?
- luminal-ai/luminal has 2.9k stars on GitHub.
- Where can I find luminal?
- luminal-ai/luminal is on GitHub at https://github.com/luminal-ai/luminal.