onnxruntime

1a9f25ff - Docs: update CUDA EP package guidance for CUDA 13 default (1.27+) and CUDA 12.8 legacy (#29781)

Docs: update CUDA EP package guidance for CUDA 13 default (1.27+) and CUDA 12.8 legacy (#29781) The CUDA Execution Provider docs were still centered on CUDA 12 defaults, which made it unclear which `onnxruntime-gpu`/NuGet GPU packages map to CUDA 13-era releases. This updates the requirements guidance to reflect the package baseline shift starting in 1.27. - **Requirements baseline update** - Updated CUDA minor-version compatibility text to explicitly cover CUDA **12.8** and **13.0** package builds. - Replaced the outdated “1.19 => CUDA 12 default” note with current guidance: - **1.27+** default GPU packages use **CUDA 13.0** (PyPI + NuGet) - Older default GPU packages use **CUDA 12.8** - **Version matrix coverage** - Added a new **CUDA 13.x** section for ORT **1.27.x–1.29.x**. - Added a **CUDA 12.8** row for ORT **1.21.x–1.26.x** in the CUDA 12.x table. - **Doc clarity for package selection** - Clarifies how to choose package lines by CUDA major/minor expectations across recent ORT releases. ```markdown Note: Starting with version 1.27, GPU packages published to PyPI (`onnxruntime-gpu`) and NuGet (`Microsoft.ML.OnnxRuntime.Gpu`) are built with **CUDA 13.0** by default. Older GPU package versions are built with **CUDA 12.8** by default. ``` --------- Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com>