From 91a0b708f1b5ceeda24c5e37db8452e1a500f68a Mon Sep 17 00:00:00 2001 From: "Gordon Lam (SH)" Date: Mon, 13 Apr 2026 16:06:02 +0800 Subject: [PATCH] Add Olive link and auto-selection info to existing optimization note --- docs/new-windows-ml/overview.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/docs/new-windows-ml/overview.md b/docs/new-windows-ml/overview.md index e6c3ebee..f64b58d3 100644 --- a/docs/new-windows-ml/overview.md +++ b/docs/new-windows-ml/overview.md @@ -52,7 +52,7 @@ This eliminates the need to: - Handle execution provider updates manually > [!NOTE] -> You're still responsible for optimizing your models for different hardware. Windows ML handles execution provider distribution, not model optimization. See [AI Toolkit](https://code.visualstudio.com/docs/intelligentapps/modelconversion) and the [ONNX Runtime Tutorials](https://onnxruntime.ai/docs/tutorials/) for more info on optimization. +> Windows ML automatically selects the best execution provider for your hardware, but you're still responsible for optimizing your models. Combining hardware-aware model optimization with the right execution provider yields the best inference performance. See [AI Toolkit](https://code.visualstudio.com/docs/intelligentapps/modelconversion), [Olive](https://microsoft.github.io/Olive/), and the [ONNX Runtime Tutorials](https://onnxruntime.ai/docs/tutorials/) for more info on model optimization. ## Performance optimization