---
license: apache-2.0
pipeline_tag: text-to-image
---

# ***ControlNet Depth SDXL, support zoe, midias***
![images](./masonry.webp)

# Example
![images_0)](./000000_depth_concat.webp)

![images_1)](./000003_depth_concat.webp)

![images_2)](./000005_depth_concat.webp)

![images_3)](./000006_depth_concat.webp)

![images_4)](./000014_depth_concat.webp)

![images_5)](./000015_depth_concat.webp)

![images_6)](./000022_depth_concat.webp)

![images_7)](./000023_depth_concat.webp)

![images_8)](./000026_depth_concat.webp)

![images_9)](./000031_depth_concat.webp)

# How to use it
```python
from diffusers import ControlNetModel, StableDiffusionXLControlNetPipeline, AutoencoderKL
from diffusers import DDIMScheduler, EulerAncestralDiscreteScheduler
from PIL import Image
import torch
import random
import numpy as np
import cv2


from controlnet_aux import MidasDetector, ZoeDetector


processor_zoe = ZoeDetector.from_pretrained("lllyasviel/Annotators")
processor_midas = MidasDetector.from_pretrained("lllyasviel/Annotators")


controlnet_conditioning_scale = 1.0  
prompt = "your prompt, the longer the better, you can describe it as detail as possible"
negative_prompt = 'longbody, lowres, bad anatomy, bad hands, missing fingers, extra digit, fewer digits, cropped, worst quality, low quality'

eulera_scheduler = EulerAncestralDiscreteScheduler.from_pretrained("stabilityai/stable-diffusion-xl-base-1.0", subfolder="scheduler")


controlnet = ControlNetModel.from_pretrained(
    "xinsir/controlnet-depth-sdxl-1.0",
    torch_dtype=torch.float16
)

# when test with other base model, you need to change the vae also.
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,
    safety_checker=None,
    torch_dtype=torch.float16,
    scheduler=eulera_scheduler,
)

# need to resize the image resolution to 1024 * 1024 or same bucket resolution to get the best performance

img = cv2.imread("your original image path")

if random.random() > 0.5:
    controlnet_img = processor_zoe(img, output_type='cv2')
else:
    controlnet_img = processor_midas(img, output_type='cv2')


height, width, _  = controlnet_img.shape
ratio = np.sqrt(1024. * 1024. / (width * height))
new_width, new_height = int(width * ratio), int(height * ratio)
controlnet_img = cv2.resize(controlnet_img, (new_width, new_height))
controlnet_img = Image.fromarray(controlnet_img)


images = pipe(
    prompt,
    negative_prompt=negative_prompt,
    image=controlnet_img,
    controlnet_conditioning_scale=controlnet_conditioning_scale,
    width=new_width,
    height=new_height,
    num_inference_steps=30,
    ).images

images[0].save(f"your image save path, png format is usually better than jpg or webp in terms of image quality but got much bigger")
```