AI drawing (4) how to write a good prompt for stable diffusion II

When drawing, a lot of styles or details will be added to describe you, and of course some professional vocabulary, such as wide angle, photo, 3D model, etc., will be introduced in detail below.

stylistic cue words

Photo (photoshop), 3D modeling (3d model), decorative art (art deco), stone sculpture (stone sculpture) and other styles.

cue word for artist style


The prompt contains the name of an artist whose style the AI ​​will try to imitate.
For example, if you add "artwork by AIfons Maria Mucha", add the style prompt of "masterpiece" (masterpiece)

Prompt words for creating specific anime characters


Just add the name and title of the work. As long as the name of the famous anime character appears, the original style will be generated without specifying the physical characteristics. For example, asuka_langley_souryuu from neon_genesis_evangelion is designated asuka_langley_souryuu from neon_genesis_evangelion

Negative Prompts


Purpose: to avoid bad outcomes


1. Add some common "bad" picture features


Deformed, lowres, bad anatomy, text
, error, extra digit, fewer digits, cropped, worst quality, low quality, normal quality, jpeg artifacts, signature , watermark, username, blurry, artist nam


2. Things you don’t want to see can also be added


Such as: I don’t want to see elements of nudity, weapons, blood, and novelties appear
nsfw, weapon, blood, guro, lowres, bad anatomy, text, error, extra digit, fewer digits, cropped, worst quality, low quality, normal quality, jpeg artifacts, signature, watermark, username, blurry, artist name
Note: The more prompts the better, the principle of positive prompts should not exceed 150 tokens; however, negative prompts can try to avoid unwanted stuffed in.

Attention/emphasis


The attention brackets are to increase the attention of the prompt word on the picture. That is, the brackets are the key elements you want to emphasize. The more brackets there are, the more the generated results will match the prompt words in the brackets.


1. ( ) The more layers of parentheses, the higher the weight (1 times without parentheses, multiplied by 1.1 times for each layer of parentheses added) For example, if you
emphasize "blue eyes", the result will have a higher chance of being blue eyes Character
one woman,((blue eyes)), long hair.
In this, the weight of blue eyes in the picture will be higher than other words.
It can also directly state how many times the attention should be increased,
one woman,(blue eyes:1.2),long hair.


2. [ ] square brackets, the more the weight, the weaker the weight (multiplied by 1.1 times), usually used in the negative prompt word field.

Prompt editing Prompt editing


Purpose: To specify which step to switch prompt words, the brackets here have nothing to do with the weight.
The syntax is [prompt word 1: prompt word 2: number of steps to be switched]
For example, set 20 steps, count apples at the beginning, and switch to oranges when the count reaches 15 steps, where 20 steps is the number of paper iteration steps.
one woman,((blue eyes)),long hair, holding fruit, [apple: orange: 15] on table

Alternating Words


Use | to switch prompt words at each step, for example, I want to switch between "blue eyes" or "red eyes" or "black eyes" when calculating a picture One woman, [blue eyes|red eyes|black eyes
]

Composable Diffusion


Purpose: Let the AI ​​generate two different objects according to the weight. For example, to generate oranges and apples, connect One woman, orange AND apple, and on table
with uppercase AND in the middle and add a number to control the weight. The initial value is 1, and the value is invalid if it is lower than 0.1. One woman, orange :1.5 AND apple:2.5 This syntax is suitable for use with ControlNet and Latent Couple, which can accurately control the position of objects in the diagram.



Prompt example

Negative cues are somewhat generic.

1. cute dog

Prompt:corgi dogs running on grass field

Negative Prompt:lowres, bad anatomy, text, error, extra digit, fewer digits, cropped, worst quality, low quality, normal quality, jpeg artifacts,signature, watermark, username, blurry, artist name

2. High-tech future city

Prompt:city future, 8k, exploration, cinematic, realistic, unreal engine, hyper detailed, volumetric light, moody cinematic epic concept art, realistic matte painting, hyper photorealistic

Negative Prompt:lowres, bad anatomy, text, error, extra digit, fewer digits, cropped, worst quality, low quality, normal quality, jpeg artifacts,signature, watermark, username, blurry, artist name

3. Futuristic technology car

Prompt:concept art,car, aerodynamic, future

Negative Prompt:lowres, bad anatomy, text, error, extra digit, fewer digits, cropped, worst quality, low quality, normal quality, jpeg artifacts,signature, watermark, username, blurry, artist name

4. Anime girls

Prompt:jeanne d'arc from fate grand order, 1girl, (best quality), (masterpiece), (high detail), ((full face)), sharp, ((looking at viewer)), ((detailed pupils)), (thick thighs), (((full body))), (large breasts)

Negative Prompt:lowres, bad anatomy, bad hands, text, error, missing fingers, extra digit, fewer digits, cropped, worst quality, low quality, normal quality, jpeg artifacts,signature, watermark, username, blurry, artist name

epilogue

So, today we introduce here. Interested friends can follow us, we will continue to update this series of tutorials.

If you think it is good and helpful to you, you can follow me, and more content will be released in the future. You can also join our planet, pay attention to the official account, reply to "planet", get the invitation code, and you can join. If you don't know what our planet is, you can read this article, which has a detailed introduction.

Empowering AI tools to achieve efficient cash realization and early leave from work

 I am discussing interesting topics with my friends at "AI High Efficiency [Work, Study]", why don't you come together?
https://t.zsxq.com/11KVfN0RChttps://t.zsxq.com/11KVfN0RC

The public account @AI宅player No. 2 player is an obsessed practitioner who relies on AI to help study and work. There are 20,000+ fans on the whole network, and the AI ​​that Xiaobai can also play is continuously updated, allowing you to leave work 2 hours earlier.

What's on the planet:
e-competition column, deep learning column, AI-assisted learning column, AI tool monetization column, etc.

Suitable for the crowd:
Students in school: I hope my experience can help you solve the problems you encounter.
Developers: Share your own experience, and you can broaden your path and learn more.
Hope to make money with side jobs: We use AI The way of cashing out can help you make money with sideline business

In fact, rigid standards cannot limit us with infinite possibilities, so ah! Come on guys!

Guess you like

Origin blog.csdn.net/qq_48764574/article/details/132253210