Ultralytics Inference

Deploying YOLO to production or edge devices with Python is bulky and slow. Ultralytics Inference is the official high-performance YOLO inference engine in Rust.

Built on ONNX Runtime with CUDA and CoreML backends, it also compiles to WebGPU/WASM and ships a CLI.

Features

Rust:Fast and lean.

Backends:CUDA and CoreML.

Browser:WebGPU and WASM.

Tasks:Detect, classify, depth.