Note
Access to this page requires authorization. You can try signing in or changing directories.
Access to this page requires authorization. You can try changing directories.
Windows ML CLI is a command-line tool for building portable, performant, and high-quality AI models for Windows ML. You can start from a source model on Hugging Face or from your own pipeline, then produce a hardware-optimized artifact in a reproducible workflow.
Windows ML CLI handles conversion, graph optimization, and compilation across AMD, Intel, NVIDIA, and Qualcomm targets. You can also use it in CI/CD pipelines to validate models and prepare them for release.
What you can do
- Build once, run across hardware - Use primitive commands such as
export,analyze,optimize,quantize, andcompile, or usewinml buildto generate a workflow that produces portable models for Windows hardware. - Inspect each stage - Review operator compatibility, shape mismatches, graph optimizations, and execution-provider-aware tuning throughout the pipeline.
- Automate model preparation - Integrate CLI-driven workflows with your existing build and release automation.
What you get out of the box
- Windows ML execution provider support - Use the same commands across supported execution providers.
- Curated model catalog - Start from a verified list of supported models that run across Windows ML execution providers.
- Bring your own ONNX model - Analyze and optimize an existing ONNX model, even if you did not convert it from PyTorch.