webonnx/wonnx

Health v2 · maintenance only

Flags: archived no_license

Adoption not part of the score

1754 stars · 71 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

Wonnx is a GPU-accelerated ONNX inference runtime written entirely in Rust, built on wgpu and usable natively or in the browser via WebGPU and WebAssembly. It ships as a Rust crate, a Python package, an npm WASM module, and a CLI tool for running ONNX models on any GPU backend (Vulkan, Metal, DX12, GLES3).

Use cases

• run onnx models on the gpu from rust
• execute onnx inference in the browser with webgpu
• accelerate neural network inference with vulkan or metal
• run onnx models from python on the gpu
• inspect and run onnx models from the command line
• deploy wasm-based ml inference client-side

When to choose

• you need GPU-accelerated ONNX inference in Rust without native ML framework dependencies
• you want to run ONNX models in the browser via WebAssembly and WebGPU
• you need a portable inference runtime across Vulkan, Metal, and DX12 backends

When to avoid

• you need full ONNX opset coverage or exotic operators that GPU runtimes may not support
• you need CPU-only inference on machines without Vulkan/Metal/DX12
• you need mature production support, quantization tooling, or training capabilities

Facets

library · maturity active

machine-learning llm-inference gpu-computing cli sdk machine-learning deep-learning gpu-computing web-development developer-tools rust python wasm cross-platform cli onnx webgpu wgpu inference-runtime webassembly neural-networks web-server

Member repositories

For agents

markdown · JSON · MCP: product_card(name="webonnx/wonnx")