Stable Diffusion model online use address:
https://inscode.csdn.net/@inscode/Stable-Diffusion
click me to go directly
Just click the link above, if you open your homepage, you may not be able to find this model.
1. Basic introduction
1. Background introduction
InsCode is a comprehensive code development platform that integrates online IDE, online AI programming, online large model training and SD model usage. Whether you are a junior software engineer or an AI model enthusiast, InsCode can help you quickly write code and run projects. Even here you can buy computing power with one click, train large models, and develop your own AI applications.
2. Template introduction
This environment is mainly used for learning and using Stable Diffusion. Relevant software and component libraries have been installed, and you can directly start Stable Diffusion WebUI for creation
Stable Diffusion is one of the most popular AI painting tools, it is a free and open source project. Through Stable Diffusion, you can easily generate corresponding pictures through text descriptions.
Stable Diffusion WebUI encapsulates the Stable Diffusion model, provides a more concise and easy-to-operate interface, and can also obtain more capabilities through plug-ins and other methods. Currently, Stable Diffusion WebUI is the most used, because it has a friendly interface and can easily adjust various parameters to generate high-quality images.
2. Use process
1. Enter the website
Click the link, go to the official website, and then click the button "Run and Use"
2. Buy a GPU
Fill in and select different configurations respectively, as follows: "Title", "Billing Method", "Purchase Time", "GPU Model", "CPU Quantity", different individual configurations correspond to different configuration fees, and the fees can be viewed on the page . Once selected, click the button "Buy GPU and create project"
3. Prepare to enter
Enter the workbench, select computing resources, recruit purchased ones, and select Stable Diffusion WebUI
4. Understand the working interface
It should look like this after going in:
Some comments I made:
You might be better off looking at this.
3. Model related version and parameter configuration:
Steps(采样迭代步数): 20,
Sampler(采样方法): Euler a,
CFG scale: 7,
Size: 512x512,
Model hash: 74c61c3a52,
Model(模型): GuoFeng3,
Version: v1.2.0
Version:v1.2.0
随机种子(seed):-1
It is not well understood here, so we will use its default properties.
4. Image generation prompts and reverse prompts
Before we do that, let's explain, generate prompts and reverse prompts:
- Generate prompt words: image generation prompt words are a group of words or phrases, which are provided to the AI model as a guide for generating images and generated.
- Negative cues: Negative cues are words or phrases that are opposite to the desired outcome and are used to indicate that the machine should not generate image features associated with them.
It's exactly where I marked it in the picture! ! !
After setting these, we just click the generate button next to it, let's start to look forward to what nice pictures it will draw?
5. Seeds and corresponding pictures:
Since we are using a random seed, the random seed is used to initialize the random state of the model. Using the same seed will ensure the same plot results under the same conditions. Different drawing effects can be obtained by changing the seed value.
1. Picture of the first chapter:
Prompt: Beautiful girl cultivating immortals, in ancient costume, holding a sword, looking into the distance
negataive prompt: Bright
This is what I want to know about my sword, the next one! ! !
2. The picture of the second chapter
prompt: Rich colors, mature charm, light yellow coat, white butterfly, delicate face, beautiful woman, close-up upper body, light blue background, beautiful woman negataive prompt: none This picture is quite good, give me a thumbs
up
! ! !
3. The third chapter picture
Prompt: In the world of cultivating immortals, female killer, wearing a veil to see her face clearly, beautiful woman, holding a dagger
Negataive prompt: None
Admit it, look up to him! ! !
4. The fourth picture
prompt: What do you think is the most beautiful girl in ancient costume
? Negataive prompt: None.
It’s okay. I’m not good at evaluating this, it’s subjective! ! !
5. Picture of Chapter 5
Prompt: highest quality, beautiful eyes, long dress, broken glass, golden petals, romantic
negataive prompt: None
This painting is pretty good, but my flower is missing! ! !
6. Summary
For this experiment, I briefly summarize:
Overall, the resulting pictures met my expectations. InsCode Stable Diffusion is very good, and it is an AI drawing model worth promoting. After all, it has reached this point at the beginning, and I personally feel that it is still very good, and other things have become better, and everyone is welcome to use it.