oxinfer 0.0.5

Local LLM inference in pure Rust (Gemma 4, Qwen 3.5/3.8): SIMD CPU kernels everywhere, GPUs detected at run time, and a WebAssembly + WebGPU browser build.

Documentation

1
2

# The kernels and loaders favour long, flat lines; wrapping them at 100 hurt readability.
max_width = 140