[FEATURE] Model browser should show npu hostable models

[aclinick]

Microsoft just released deepseek with npu support https://blogs.windows.com/windowsdeveloper/2025/01/29/running-distilled-deepseek-r1-models-locally-on-copilot-pcs-powered-by-windows-copilot-runtime/ and https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B

This seems like a perfect candidate for support in the dev gallery (which is quite excellent btw great job) since it really makes the npu on a copilot+ pc shine.

If you search for DeepSeek-R1-Distill-Qwen-1.5B in dev gallery you get responses but none of them are the NPU optimized version.

It would be great if you could filter the search to only show models that can run on your gpu, or the npu on your system. This is likely the only feature i miss from LM studio

[nmetulev]

Thanks @aclinick. We are on track to enable this in the next few weeks when the onnxruntimegenai.qnn version that supports these models is available more broadly.