Issue Installing Intel Ipex Extension on GPU Instance

[Nate_]

Hi. I'm currently having an issue installing Ipex/pytorch extensions for GPU that I wasn't having prior. Using version 2.1 is the CPU version according to the GitHub page, so that won't solve this. 

The following code is how I have been installing it in the past, and as far as I can see on the intel Github it is the most current version listed. As recently as last week this was still behaving as expected.

code:

python -m pip install torch==2.0.1a0 torchvision==0.15.2a0 intel_extension_for_pytorch==2.0.110+xpu -f https://developer.intel.com/ipex-whl-stable-xpu

I saw a note that said temporarily it needs to be torch 2.1 but that doesn't solve the issue. It is still giving me an error of:

Defaulting to user installation because normal site-packages is not writeable
Looking in links: https://developer.intel.com/ipex-whl-stable-xpu

ERROR: Could not find a version that satisfies the requirement intel_extension_for_pytorch==2.0.110+xpu (from versions: 2.0.0, 2.0.100, 2.1.0)
ERROR: No matching distribution found for intel_extension_for_pytorch==2.0.110+xpu

The instance is bare metal with the following specs: 
Intel® Max Series GPU (PVC) on 4th Gen Intel® Xeon® processors – 1100 series (4x)
Ubuntu

If someone can let me know what has changed, so I can correct my approach that would be appreciated.  Thank you! 

[Hairul_Intel]

Hi Nate_,

Thank you for reaching out to us.

Please have a try updating your pip to the latest version and try running the command again.

python3 -m pip install --upgrade pip

If the error persists, please provide the following additional information for further investigation from our side.

Instance ID:

Reservations details

Start time:

End time:

Regards,

Hairul

[Nate_ · Accepted solution]

Hi. That didn't fix it, but I was given a solution by an Intel Engineer that I will share here in case this helps anyone else.

The following solved it for me where [environment_name] is obviously what you name your environment.

Solution:

conda create -n environment_name python==3.9

conda activate environment_name

python -m pip install torch==2.0.1a0 torchvision==0.15.2a0 intel_extension_for_pytorch==2.0.110+xpu --extra-index-url https://pytorch-extension.intel.com/release-whl/stable/xpu/us/

I am unsure what changed that pip install is no longer sufficient, but this solved it for me. Thank you for your help!

Best,
Nate

[Hairul_Intel]

Hi Nate_,

Glad to know that your issue has been resolved and we appreciate you sharing the solution with the community.

This thread will no longer be monitored since this issue has been resolved. If you need any additional information from Intel, please submit a new question.

Regards,

Hairul