Skip to content

Latest commit

 

History

History
37 lines (24 loc) · 1.08 KB

File metadata and controls

37 lines (24 loc) · 1.08 KB

Installation for Intel GPUs

First of all, I have confirmed that this software works on Intel Max 1100 on Intel Tiber AI Cloud. However, I don't own any Intel GPU hardware myself, so I haven’t set it up myself. (The setup has already been completed on Intel Tiber AI Cloud.)

There are two main differences compared to a standard installation.

  1. Install Intel GPUs Driver
  2. Install PyTorch for Intel GPUs (xpu)

1. Install Intel GPU Driver

See https://www.intel.com/content/www/us/en/developer/articles/tool/pytorch-prerequisites-for-intel-gpu.html

There is no need to install "Intel® Deep Learning Essentials".

2. Install PyTorch for Intel GPUs (xpu)

Use requirements-torch-xpu.txt instead of requirements-torch.txt.

pip3 install -r requirements-torch-xpu.txt

If you are using nunif-windows-package, change the following line in update.bat.

python -m pip install --no-cache-dir --upgrade -r "%NUNIF_DIR%\requirements-torch.txt"

to

python -m pip install --no-cache-dir --upgrade -r "%NUNIF_DIR%\requirements-torch-xpu.txt"

After that, run update.bat.