> The issue is, no tech company is interested in adding D3D or Vulkan backend i'm using Apache TVM right now to run LLMs via Vulkan on Intel Arc cards. Intel

> The issue is, no tech company is interested in adding D3D or Vulkan backend i'm using Apache TVM right now to run LLMs via Vulkan on Intel Arc cards. Intel has their own pytorch wheels here for their GPUs: https://developer.intel.com/ipex-whl-stable-xpu https://developer.intel.com/ipex-whl-stable-xpu