【Stable Diffusion】キレイなお姉さんを描くプロンプト5つの基本

13 May 202313:11

TLDRThe video script discusses the process of creating beautiful illustrations using AI with a short and effective prompt. It introduces a method for structuring prompts into five groups: character, clothing, action, location, and enhancing keywords. The video demonstrates the use of Stable Diffusion with an NVIDIA RTX 3060 and shares tips for beginners to start by copying prompts. It also explains the importance of understanding the role of each element in the prompt to achieve desired results and suggests seeking assistance from AI like ChatGPT for English terminology.


  • 🎨 The video discusses techniques for creating beautiful illustrations using AI with short and effective prompts.
  • 🖌️ Viewers are encouraged to start by copying provided prompts to understand the process and then gradually make their own variations.
  • 📝 The importance of understanding the basic structure of prompts is emphasized, which includes five main groups: subject, clothing, action, location, and quality-enhancing keywords.
  • 👗 The video provides examples of how to describe clothing in prompts, such as 'long hair', 'ribbon', 'school uniform', and 'casual clothes'.
  • 🎭 The script explains the role of negative prompts in refining the output and maintaining the desired quality of the illustrations.
  • 🔢 The use of parameters like sampling method, sampling steps, and cfg scale in the AI's interface are briefly introduced to give control over the illustration's quality.
  • 🌟 The video demonstrates how to create a variety of illustrations by changing elements in the prompts, such as character type, clothing, actions, and settings.
  • 👩‍🎤 The process of using AI to draw is likened to giving instructions to an AI, with keywords acting as those instructions.
  • 🌈 The importance of learning English terminology for specific clothing and accessories is highlighted to improve the accuracy of prompts.
  • 📚 The video encourages viewers to use ChatGPT for learning English terms related to prompts to enhance their AI drawing capabilities.
  • 🎁 The script concludes by encouraging viewers to explore further and apply what they've learned to create diverse and high-quality illustrations.

Q & A

  • What was the main topic of the previous video mentioned in the script?

    -The main topic of the previous video was using NVIDIA's RTX 3060 to run stable, diffusion for creating images.

  • What is the purpose of the prompt in the context of the video?

    -The purpose of the prompt is to provide instructions to the AI in order to generate specific images. It is used to guide the AI in creating the desired artwork by specifying details such as the subject, clothing, actions, and setting.

  • How does the video suggest a beginner should start with creating images using AI?

    -The video suggests that beginners should start by copying and pasting the provided prompts to generate images and then gradually learn to modify them as they become more familiar with the process.

  • What are the five groups of elements the video suggests considering when writing a prompt?

    -The five groups of elements are: 1) Who is the subject (e.g., male, female, human, animal), 2) What is the subject wearing (e.g., long hair, ribbon, school uniform), 3) What is the subject doing (e.g., sitting, smiling), 4) Where is the subject (e.g., classroom, beach), and 5) Standard keywords to make the image aesthetically pleasing.

  • What is the significance of the 'standard keywords' mentioned in the script?

    -Standard keywords are phrases or terms that are added to the prompt to enhance the quality of the generated image. They act as a kind of 'charm' or 'spell' that helps to ensure the image turns out well without needing to understand all the details.

  • How does the video describe the process of selecting a sampling method for creating images?

    -The video describes the process as choosing a mechanism for assembling the image. DTM +2m is suggested as a good choice for the sampling method.

  • What role does the 'cfg scale' parameter play in the image generation process?

    -The 'cfg scale' parameter determines how closely the AI follows the instructions provided in the prompt. A lower value may result in the image not adhering to the prompt, while a higher value may cause the image to be too focused on the prompt and potentially become distorted.

  • What is the purpose of the 'seed value' in the image generation process?

    -The 'seed value' is used for random selection in the image generation process. By keeping the prompt the same but changing the seed value, different images can be generated, adding variety to the outputs.

  • How does the video demonstrate changing the theme of the generated images?

    -The video demonstrates changing the theme by altering the subject, clothing, actions, and settings in the prompts. For example, it shows how to change from a school girl theme to a bunny girl theme, or from a casual outfit to a business suit.

  • What advice does the video give for users who are unsure about specific English terms for clothing or items?

    -The video advises users to consult AI tools like ChatGPT for translations and correct terminology. This helps in creating more accurate and diverse prompts for generating images.

  • What is the main takeaway from the video regarding the use of prompts for AI image generation?

    -The main takeaway is that understanding the basic structure of prompts and how to modify them can greatly enhance the ability to generate a variety of images. It emphasizes the importance of experimenting with different elements in the prompts to create unique and desired outputs.



🎨 Introduction to Stable Diffusion with RTX 3060

This paragraph introduces the use of NVIDIA's RTX 3060 to run Stable Diffusion, a machine learning model for generating images. It discusses the importance of crafting short and effective prompts to produce beautiful images. The video provides examples of prompts used, which are also listed in the description section for beginners to copy and try out. The content emphasizes the process of learning to use Stable Diffusion by starting with basic prompts and gradually making changes to create a variety of images. It also touches on the concept of using AI for art creation and the role of keywords in instructing the AI.


🌟 Enhancing Prompts for Better Art

The second paragraph delves into the strategy of enhancing prompts to achieve a more desirable outcome in art creation. It explains the importance of adding specific prompts to prevent certain tendencies, such as making a character appear too masculine, and suggests ways to make the art feel more feminine. The paragraph also introduces the concept of using a 'mystery incantation' or a set of keywords that contribute to the quality of the final image. It highlights the role of randomness in the creation process due to the seed value, which can lead to different outcomes even with the same prompt.


👗 Exploring Various Themes and Costumes

This paragraph explores different themes and costumes for art creation using Stable Diffusion. It discusses the process of selecting and refining prompts to generate images of characters in various attires, such as bunny girl outfits, school uniforms, and nurse costumes. The content emphasizes the subtle differences between 'girl' and 'older sister' themes and provides tips on how to control age in the art. It also touches on the challenge of understanding and using English keywords effectively, suggesting the use of AI like ChatGPT for assistance in such cases.




NVIDIA RTX 3060 is a high-performance graphics processing unit (GPU) that is central to the video's theme of utilizing technology for AI-based image generation. The GPU is used to power the stable, diffusion process, which is a method of creating images using AI algorithms. In the context of the video, it is the hardware that enables the user to run the AI models that generate the images.


A prompt, in the context of AI image generation, refers to a set of instructions or keywords that guide the AI in producing a specific output. It is a critical component as it directly influences the final image created by the AI. The video emphasizes the importance of crafting effective prompts to achieve desired results in image generation.

💡Stable Diffusion

Stable Diffusion is an AI model used for generating images from textual descriptions. It is a type of deep learning algorithm that has been trained on large datasets to understand and produce visual content based on textual prompts. In the video, Stable Diffusion is the technology that the user interacts with to create images, showcasing its role in the realm of AI and art.

💡AI Art Generation

AI Art Generation refers to the process of creating visual art through artificial intelligence, where AI models are fed prompts to produce images, paintings, or other forms of visual content. This concept is the core theme of the video, as it explores the use of AI and specific technologies like Stable Diffusion to generate images based on user input.


In the context of AI image generation, parameters are the adjustable settings within the AI model that control the output's characteristics. These can include aspects like image resolution, color saturation, and the level of detail. Parameters are crucial as they allow users to refine the AI's output to align with their creative vision.

💡Sampling Method

Sampling Method refers to the technique used by AI models to select and combine elements from the training data to create the final output. In AI Art Generation, different sampling methods can lead to varying levels of detail, style, and overall quality of the generated images. It is a key concept in understanding how AI translates prompts into visual content.

💡Negative Prompt

A negative prompt is a set of instructions provided to an AI model to avoid certain elements or characteristics in the generated image. It is used to guide the AI away from undesired outcomes and towards the desired result. In the video, negative prompts are used to ensure the AI focuses on creating images that match the user's creative intent.

💡Seed Value

A seed value is a starting point used in random number generation algorithms to ensure that the AI can produce different outputs even with the same prompt. It introduces variability in the AI's responses, allowing for the creation of unique images without altering the core prompt.


Fashion, in the context of the video, refers to the clothing and styles assigned to the characters in the AI-generated images. It is an essential aspect of the creative process, as it helps define the characters' personalities and the overall mood of the image.


A character in AI Art Generation is the central figure or subject of the image being created. The video focuses on generating human or anthropomorphic characters with specific attributes, expressions, and actions as defined by the user's prompts.


Setting in the context of AI-generated art refers to the background, environment, or scene in which the character exists. It adds context and depth to the image, enhancing the narrative and visual appeal. The setting is a crucial element in the creative process, as it complements the characters and actions described in the prompts.

💡Image Quality

Image quality refers to the visual clarity, detail, and aesthetic appeal of the AI-generated images. High-quality images are more realistic, have better resolution, and exhibit greater attention to detail, which is important for achieving a professional and polished look.


Introduction to using NVIDIA RTX 3060 for stable, diffusion-based image generation.

Explanation of crafting short yet effective prompts for AI image generation.

The importance of understanding the basic structure of prompts to create various images.

How to begin with simple prompts and gradually introduce complexity to generate diverse images.

The five key groups to consider when constructing prompts: subject, attire, action, location, and aesthetic enhancements.

Utilizing negative prompts to refine and improve image generation outcomes.

Web UI overview and the selection of learning models for image generation.

Parameters explanation, including sampling methods and steps for image assembly.

The significance of seed values in creating unique images with the same prompt.

Demonstration of creating an image using a girl-like character with a specific outfit and setting.

The process of generating an image with a 'cute sister' theme and aesthetic lighting effects.

Exploration of generating images with different themes, such as a bunny girl or a heroic outfit.

Adjusting and experimenting with prompt elements to achieve desired image variations.

The role of English vocabulary in crafting prompts and utilizing AI tools for language support.

Encouragement for viewers to explore and practice with prompts to enhance their understanding and creativity.

Concluding remarks on the growth and potential of AI image generation through learning and experimentation.