ControlNet Review

ControlNet Soars to New Heights!! 😍😍😍😍 Getting better and better to use!! 🚀🚀🚀 As one of the important plugins for Stable Diffusion, ControlNet has evolved from an important feature to a main feature. Many functions cannot be achieved without ControlNet, and many functions of CN have covered the original functions of SD, isn't it soaring! 💪💪💪

At this pace 🔥🔥🔥, creating a webui process centered around CN is not impossible.

Major Update V1.1.173🆕🆕 brought by author Mikubill this time
[Reference-Only Control]
Even without a model, 💥💥💥 just update the plugin to the latest version, select reference-only in preprocess, and you can experience it.

This update has certain similarities with the image generation function of SD, but solves the problem of potential failure in the original SD functions.

[Reference-Only] VS [Original Image Generation]
The original one more refers to the style and color of the original image, 🎨 controls the similarity of the generated image according to the degree of redrawing, and if the degree of redrawing is too low, it can be completely consistent, and if it is too high, it may be very different.

[Original Image Generation] Advanced Gameplay:
Want to generate an image based on the original image A, the original function is quitetroublesome:
Paste the original image A and a white image of the same size together, use the inpaint function, paint the white image black, and then generate a similar image based on the original image A. However, it is easy to fail andtroublesome.

[Reference-Only] greatly simplifies the process of padding images, similar to MJ's padding, and is more controllable in various aspects. Pairing with Multi ControlNet is simply too cool!!! 💯💯💯

ControlNet Precision Control AI Painting Tutorial 🎨

In recent years, AI painting technology has gradually entered our lives, giving rise to a series of stunning artworks. However, it is not easy to generate perfect and satisfactory images with AI painting. Multiple attempts and adjustments of keywords, sometimes still difficult to achieve the expected effect. This has always been a difficulty in the field of AI painting.

However, recently, an AI painting plugin named "ControlNet" has emerged, almost perfectly solving the above problems, making the pain points of AI painting a thing of the past. The ControlNet plugin has the ability to precisely control the scene, making the generated painting works like a miracle of finger dancing.

The painting mode of ControlNet is very simple. The user provides a reference image first, and then the plugin generates a new preprocessed image according to this image. Then, the AI will draw a perfect painting work based on these two images. Of course, the user can also choose to turn off the pre-processing function and directly use the user's own processed image as the pre-processing image, and the AI drawing result will be closer to the user's expectations.

Let's take the example of four beautiful girls to illustrate the magic of ControlNet. Originally, this was a common group photo of four best friends on the beach. But with the help of ControlNet, this photo instantly changed dramatically, as if it had come to a beautiful anime world. Not only did the style change, but the poses of the characters also remained true to life. These four beautiful girls seem to have realized their dream of freely traversing the anime world.

📷 Image Display: (Insert the emoji of a beautiful girl)

It is ControlNet that provides a new solution for the development of AI painting. It is not just a plugin, but also brings infinite creative possibilities to users. Whether pursuing anime style or other styles, ControlNet can accurately capture user needs and generate corresponding painting works.

For art enthusiasts, ControlNet will be an indispensable tool. It not only provides a convenient operation method, but also allows users to participate in the entire process of art creation through precise control of the AI painting process.

In summary, ControlNet is a plugin that can precisely control AI painting, changing the pain points of AI painting, providing more creative possibilities, and allowing users to easily realize their artistic dreams. Whether pursuing anime style or other styles, ControlNet can meet your needs and make your painting works more outstanding!

If you are interested in AI painting, why not try the ControlNet plugin? I believe it will bring you surprises. Let's explore the infinite possibilities of AI painting together! ✍️

ControlNet Installation Tutorial

ControlNet and other WebUI extensions based on injection for SD.

This extension is suitable for AUTOMATIC1111's Stable Diffusion Web UI, allowing the Web UI to add ControlNet to the original stable diffusion model to generate images. The addition is immediate and does not require merging.

Installation

1. Open the "Extensions" tab.
2. Open the "Install from URL" tab in the tab.
3. Enterhttps://github.com/Mikubill/sd-webui-controlnet.git"URL of the extension's git repository".
4. Click the "Install" button.
5. Wait for 5 seconds, you will see the message "Installed to stable-diffusion-webui\extensions\sd-webui-controlnet. Restart the installed tab".
6. Navigate to the "Installed" tab, click "Check for Updates," then click "Apply and Restart UI." (Next time, you can use these buttons to update ControlNet.)
7. Fully restart A1111 webui, including your terminal. (If you're unsure what a "terminal" is, restarting your computer will achieve the same effect.)
8. Download the model (see below).
9. After placing the model in the correct folder, you may need to refresh to see the model. The refresh button is located to the right of the "Model" dropdown list.

Download model

Currently, all 14 models of ControlNet 1.1 are in beta testing phase.

Download models from ControlNet 1.1:

https://huggingface.co/lllyasviel/ControlNet-v1-1/tree/main

High-speed download available here in China:

"All ControlNet plugin models":https://pan.quark.cn/s/f1b98f5e0870

You need to download model files ending with ".pth".

Place the model in "stable-diffusion-webui\extensions\sd-webui-controlnet\models". You only need to download the "pth" files.

Do not right-click on the filenames on the HuggingFace website to download. Some users right-click on these HuggingFace HTML websites and save these HTML pages as PTH/YAML files. They did not download the correct files. Please click the small download arrow "↓" icon in HuggingFace to download.

<