FICHA · MANJARO

onnxruntime-opt-cuda

Cross-platform, high performance scoring engine for ML models (with CUDA and AVX2 CPU optimizations)

  • Runtime
  • RUNTIME
  • MACHINE-LEARNING
  • ONNX
  • Dependency only
official+codex · reviewed · May 28, 2026 description in en

Description

Provides ONNX Runtime with CUDA support and AVX2 CPU optimizations. It targets systems that can benefit from NVIDIA GPU inference while also using optimized CPU code paths.

Use it when deployment requires this optimized build. Compatibility depends on GPU drivers, CUDA support, CPU features, model operators, and available memory.

Permissions

Permissions not analysed for this source yet.