patrickvonplaten's picture
update README (#1)
29c13b4
|
raw
history blame
2.58 kB
metadata
license: openrail
base_model: runwayml/stable-diffusion-v1-5
tags:
  - art
  - controlnet
  - stable-diffusion

Controlnet

Controlnet is an auxiliary model which augments pre-trained diffusion models with an additional conditioning.

Controlnet comes with multiple auxiliary models, each which allows a different type of conditioning

Controlnet's auxiliary models are trained with stable diffusion 1.5. Experimentally, the auxiliary models can be used with other diffusion models such as dreamboothed stable diffusion.

The auxiliary conditioning is passed directly to the diffusers pipeline. If you want to process an image to create the auxiliary conditioning, external dependencies are required.

Some of the additional conditionings can be extracted from images via additional models. We extracted these additional models from the original controlnet repo into a separate package that can be found on github.

Pose estimation

Diffusers

Install the additional controlnet models package.

$ pip install git+https://github.com/patrickvonplaten/controlnet_aux.git
from PIL import Image
from diffusers import StableDiffusionControlNetPipeline, ControlNetModel, UniPCMultistepScheduler
import torch
from controlnet_aux import OpenposeDetector

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

image = Image.open('images/pose.png')

image = openpose(image)

controlnet = ControlNetModel.from_pretrained(
    "fusing/stable-diffusion-v1-5-controlnet-openpose", torch_dtype=torch.float16
)

pipe = StableDiffusionControlNetPipeline.from_pretrained(
    "runwayml/stable-diffusion-v1-5", controlnet=controlnet, safety_checker=None, torch_dtype=torch.float16
)

pipe.scheduler = UniPCMultistepScheduler.from_config(pipe.scheduler.config)

# Remove if you do not have xformers installed
# see https://huggingface.co/docs/diffusers/v0.13.0/en/optimization/xformers#installing-xformers
# for installation instructions
pipe.enable_xformers_memory_efficient_attention()

pipe.enable_model_cpu_offload()

image = pipe("chef in the kitchen", image, num_inference_steps=20).images[0]

image.save('images/chef_pose_out.png')

pose

openpose

chef_pose_out

Training

The Openpose model was trained on 200k pose-image, caption pairs. The pose estimation images were generated with Openpose. The model was trained for 300 GPU-hours with Nvidia A100 80G using Stable Diffusion 1.5 as a base model.