In the process of answering questions about function calculation Stable Diffusion, I encountered many students who encountered difficulties in the process of installing some plug-ins. Some of them require installing some dependencies, and some require writing some code. In many cases, it only takes a few days to install a plug-in . , we have collected many plug-ins that students need, and this time we have installed all the Stable Diffusion plug-ins that are more difficult to install. You can check according to your own needs.

Deploying Stable Diffusion Tutorial Based on Function Computing: https://developer.aliyun.com/adc/scenario/b2cc0e1c3a6244e0bd9fc0f37acd5a0e

When using it, check it in the list, click Apply and quit, and the restarted WebUI will have its own plug-in.

What can these plugins do?

ADetailer Plugin: Fix Honkaiface

When using Stable Diffusion to generate characters, you often encounter the problem of face collapse. The adetailer plug-in can effectively improve the problem of face collapse.

Open ADetailer.

Find a picture of a broken face, drag it to the [Picture Information] tab, get the picture information, then click on the Vincent picture to load the picture information settings, check enable adetailer in After Detailer, and select the model name if you want to repair the face. With faces, if you want to repair real people, I feel that the third and fourth ones are more effective. This plug-in is really effective in reshaping faces. You can see the comparison information below. It can also repair multiple faces. The Adetailer plug-in works much better than simply checking “face repair”.

Deforum Plugin: Ephemeral Universe Maker

The instant universe video became popular a while ago. Videos that constantly change into new scenes through the screen are very eye-catching. This plug-in is a must-have tool for making such videos.

Using the Deforum plug-in, you can even define key frames and the first frame yourself. Pay attention to keeping the format consistent with the example. Making a hit video is not a dream.

Prompt all in one plugin: translation assistant

This plug-in is worthy of the name all in one, and it can solve the problems of translation, adding commas, multilingual conversion... in the process of writing prompt words, and it can be called an artifact.

You can choose multiple languages for input, you can choose multiple translation interfaces, prompt word input and translation, and you can also translate all English words with one click. In this way, after you get the prompt words from other places, you don’t have to look at which word is the new word. Go to software translation. You can enter directly in the input box of webui, and click on the blank space to automatically translate. You can also collect prompt words and historical prompt words.

roop plugin: change face with one click

Many face-changing techniques require a lot of training on the face to be realized. And this plug-in only needs to provide a photo, and the face can be changed in 10 seconds.

Image Browser Plugin: Image Management

Image Browser is not only an image browser, but also a powerful image manager. Precise image search combined with multi-selection operations greatly improves efficiency. The plug-in supports image search and collection, image preview and "send to". Preview and file operations can be performed based on the file tree structure.

rembg plug-in: one-click cutout artifact

Cutting out pictures is a necessity for many designers, and it is also a relatively time-consuming and labor-intensive task. The Stable Diffusion plug-in stable-diffusion-webui-rembg can realize quick cutout of pictures with one click, which saves time and effort. The effect is very good.

Tagger plug-in: Reverse prompts for images

Compared with the tab reverse tool that comes with Stable Diffusion, this plug-in reverses keywords more accurately, and the image effect generated is closer to the original image effect.

dataset tag editor plugin

Many people call training Lora's image "alchemy". This plug-in is a necessary marking tool for alchemy. The specific usage will be detailed in the alchemy tutorial later.

How to deploy Stable Diffusion based on Function Compute

Resource preparation

Register and log in to your Alibaba Cloud account [ 1]
Activate the function calculation and receive the trial quota [ 2]
Activate the file storage NAS service and receive trial credit [ 3]

Deploy Cloud Stable Diffusion

Select "AI digital painting stable-diffusion custom template" and click Create Now to start creating Stable Diffusion* https://fcnext.console.aliyun.com/applications/create *

To create Stable Diffusion, you need to activate Function Compute FC and file storage NAS. New users must receive a trial quota

Create Stable Diffusion WebUI and select the region -> Select "Drawing Type" (you can change it yourself later) -> RAM is required for the first deployment. Role APN is authorized. If there are no special requirements, you can directly default to other places.

Confirm that you are aware of the function calculation fees and important reminders generated during the use of SD, and wait for the successful deployment of the Stable Diffusion WebUI

"Initialize model management" is required to replace and manage the model. Note that the storage model starts to use file storage NAS products

Confirm to understand the cost and precautions of using NAS, confirm to start creating a model management tool - select "automatic configuration" - confirm

Visit the domain name starting with admin to enter the model management platform

Built-in model library

You can choose your favorite model to generate the style of QR code. This tutorial recommends downloading the majicMix realistic model from the model library. The model will be downloaded to your file management (NAS), which will incur a certain file storage fee. Please Be sure to pay attention. After downloading, restart Stable Diffusion to use it.

The model library function is provided by Alibaba Cloud OSS. The models/datasets/files are all from third parties, and compliance is not guaranteed. Please consider carefully before using.

If you encounter any problems in the above steps, you can check the detailed documentation:

https://developer.aliyun.com/adc/scenario/exp/b2cc0e1c3a6244e0bd9fc0f37acd5a0e?

Odyssey in the Universe, Alibaba Cloud x Semir AIGC T-shirt Design Contest

1. [Semir X Alibaba Cloud AIGC T-shirt Design Competition]

Programmers AIGC's own t-shirt, win Airpods, graphic custom t-shirts!

Quickly deploy Stable Diffusion through Function Compute FC: built-in model library + common plug-ins + ControlNet, support SDXL1.0.

Join now: https://developer.aliyun.com/adc/series/activity/aigc_design

You can win the third generation of Airpods, customized joint t-shirts, Semir luggage and other peripherals!

2. Participate in topic activities

Discuss the future development trend of AIGC together, users can communicate and share from any angle, and you can win prizes such as eye protection lamps, data cables, and silent purifiers!

Topic: "Compared to excellent fashion designers, how can AIGC break through the boundaries of design inspiration?" Is it purely mechanical language or just a little flash of inspiration? 》

https://developer.aliyun.com/ask/548537?spm=a2c6h.13148508.setting.14.4a894f0esFcznR