tracel-ai/burn
Burn is a next generation tensor library and Deep Learning Framework that doesn't compromise on flexibility, efficiency and portability.
Leverages Rust's compile-time guarantees to enable static-graph optimizations while maintaining dynamic flexibility, with a composable backend architecture supporting GPU (CUDA, ROCm, Metal, Vulkan, WebGPU) and CPU execution. Backend decorators like Autodiff and Fusion augment any backend with automatic differentiation and kernel fusion without code changes. Supports distributed training via Remote backend and seamless CPU-GPU operation routing, while unifying training and inference workflows with no code modifications needed between them.
14,588 stars and 72,188 monthly downloads. Actively maintained with 71 commits in the last 30 days.
Stars
14,588
Forks
846
Language
Rust
License
Apache-2.0
Category
Last pushed
Mar 13, 2026
Monthly downloads
72,188
Commits (30d)
71
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/tracel-ai/burn"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Compare
Related frameworks
sonos/tract
Tiny, no-nonsense, self-contained, Tensorflow and ONNX inference
pykeio/ort
Fast ML inference & training for ONNX models in Rust
robertknight/rten
ONNX neural network inference engine
elixir-nx/ortex
ONNX Runtime bindings for Elixir
tensorflow/rust
Rust language bindings for TensorFlow