How to change stable cascade prior/decoder to any sdxl unet model?

[s9anus98a]

something like this may be?

import torch
from diffusers import StableDiffusionXLPipeline,UNet2DConditionModel

unet = UNet2DConditionModel.from_pretrained("SPO-Diffusion-Models/SPO-SDXL_4k-p_10ep", subfolder="unet", torch_dtype=torch.float16,
                                            
                                            ).to('cuda')

because stable cascade less aestetics may be can change the style with other sdxl model but the image quality keep in cascade