metadata

license: apache-2.0
tags:
  - openpose
  - controlnet
  - diffusers
  - controlnet-openpose-sdxl-1.0
  - text_to_image

State of the art ControlNet-openpose-sdxl-1.0 model, not limited to anime, just for show

controlnet-openpose-sdxl-1.0

Developed by: xinsir
Model type: ControlNet_SDXL
License: apache-2.0
Finetuned from model [optional]: stabilityai/stable-diffusion-xl-base-1.0

Model Sources [optional]

Paper [optional]: https://arxiv.org/abs/2302.05543

Examples

How to Get Started with the Model

Use the code below to get started with the model.

from diffusers import ControlNetModel, StableDiffusionXLControlNetPipeline, AutoencoderKL
from diffusers import DDIMScheduler, EulerAncestralDiscreteScheduler
from controlnet_aux import OpenposeDetector
from PIL import Image
import torch
import numpy as np
import cv2



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-openpose-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,
)

processor = OpenposeDetector.from_pretrained('lllyasviel/ControlNet')


controlnet_img = cv2.imread("your image path")
controlnet_img = processor(controlnet_img, hand_and_face=False, output_type='cv2')


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

Evaluation Data

HumanArt [https://github.com/IDEA-Research/HumanArt], select 2000 images with ground truth pose annotations to generate images and calculate mAP.

Quantitative Result

metric	xinsir/controlnet-openpose-sdxl-1.0	lllyasviel/control_v11p_sd15_openpose	thibaud/controlnet-openpose-sdxl-1.0
mAP	0.357	0.326	0.209

We are the SOTA openpose model compared with other opensource models.