RuntimeError: PyTorch is not linked with support for xpu devices

[Loong]

When I run this notebook:

oneAPI-samples/AI-and-Analytics/Getting-Started-Samples/Intel_Extension_For_PyTorch_GettingStarted/ResNet50_Inference.ipynb

I met some problems:

RuntimeError: PyTorch is not linked with support for xpu devices

And here, Overview | Intel® DevCloud, you told that I can get 

GPU:

• Intel® Xeon® E-2176 P630 processors (with Intel® Graphics Technology)
• Intel® Iris® Xe MAX

But, I cannot get any Xe MAX gpu.

When I use clinfo to detect the devices, it only shows that it has an Accelerator, "Intel(R) FPGA Emulation Device" and a CPU, "Intel(R) Xeon(R) Gold 6128 CPU @ 3.40GHz".

[AthiraM_Intel]

Hi,

Thank you for posting in Intel Communities.

We are sorry to inform you that Intel® Iris® Xe Max Graphics are currently not available in DevCloud. We’re working on adding new Intel® Graphics in the future, as they become available on the market. We will make an appropriate announcement when the time comes.

Regarding the sample issue, we are checking on this internally. We will get back to you with an update.

Thanks

[Loong]

I do not notice any Xe MAX gpu on market, even though you have release them 2 years ago.

So, as you say "as they become available on the market", can I think that you will add Intel® Arc™ Graphics to DevCloud?

And when?

[vat7]

I was able to resolve the issue, using following steps:

Just add these installations

python -m pip install torch==1.13.0a0+git6c9b55e torchvision==0.14.1a0 intel-extension-for-pytorch==1.13.120+xpu --extra-index-url https://pytorch-extension.intel.com/release-whl/stable/xpu-idp/us/

Check out this for more,

https://intel.github.io/intel-extension-for-pytorch/index.html#installation?platform=gpu&version=v2.0.110%2Bxpu