fix bug in StableDiffusionXLControlNetPipeline when use guess_mode

[yiyixuxu]

This PR fix a bug in sdxl controlnet when guess_mode=True

It fix#4709 and I think it does the same thing as #4155 but we don't need to drastically change the code structure

fromdiffusersimportStableDiffusionXLControlNetPipeline, ControlNetModel, AutoencoderKLfromdiffusers.utilsimportload_imageimportnumpyasnpimporttorchimportcv2fromPILimportImageprompt="aerial view, a futuristic research complex in a bright foggy jungle, hard lighting"negative_prompt="low quality, bad quality, sketches"image=load_image(
"https://hf.co/datasets/hf-internal-testing/diffusers-images/resolve/main/sd_controlnet/hf-logo.png"
)
controlnet_conditioning_scale=0.5# recommended for good generalizationcontrolnet=ControlNetModel.from_pretrained(
"diffusers/controlnet-canny-sdxl-1.0", torch_dtype=torch.float16
)
vae=AutoencoderKL.from_pretrained("madebyollin/sdxl-vae-fp16-fix", torch_dtype=torch.float16)
pipe=StableDiffusionXLControlNetPipeline.from_pretrained(
"stabilityai/stable-diffusion-xl-base-1.0", controlnet=controlnet, vae=vae, torch_dtype=torch.float16
)
pipe.enable_model_cpu_offload()
image=np.array(image)
image=cv2.Canny(image, 100, 200)
image=image[:, :, None]
image=np.concatenate([image, image, image], axis=2)
canny_image=Image.fromarray(image)
image=pipe(
prompt, controlnet_conditioning_scale=controlnet_conditioning_scale, image=canny_image, guess_mode=True,
).images[0]
image.save("guess_mode_out.png")

[HuggingFaceDocBuilderDev]

The documentation is not available anymore as the PR was closed or merged.

[patrickvonplaten]

Cool! One quick fast/dummy test for this one?

[patrickvonplaten]

Very nice! Looks good to me - let's maybe add this also to our SDXL controlnet docs in a follow-up PR (cc @stevhliu)

Think the test is also currently failing on the CI - can we make it more deterministic?

[patrickvonplaten]

Cool! @sayakpaul maybe for a final review?

[sayakpaul]

The logic definitely reads simplistic now. Thank you!