Intel recently came out with AIPlayground2 that can run different AI image generators, image editors and chat bots like R1 locally on Intel GPUs and CPUs no cuda needed. It includes Comfyui that is used to make workflows between different
AI models as well and some small example workflows like flux1-schnell that works great for product concepts with text that is actually correct.
I was wondering if anyone else here was using it and any tips or tricks or code or workflows to share.
One bit of code I have to help is this IPEX to Cuda. It should help for custom nodes to use Intel IPEX without having to change code.