HOW TO MAKE BEAUTIFUL STABLE DIFFUSION IMAGES | Negative Prompts

Binks
13 Jan 202305:02

TLDRIn this informative video, Binks discusses the significance of negative prompts in generating images using Stable Diffusion. Binks shares his experience with Protogen version 3.4, a photorealism model available on the CivitAI page, and provides a link for viewers. He demonstrates the process of creating an image by first using a simple prompt to generate a base image and then refining the process by incorporating a negative prompt to avoid unwanted elements such as a canvas frame or disfigured subjects. Binks emphasizes the importance of crafting detailed prompts and adjusting sampling settings to achieve desired results. He encourages viewers to experiment with the AI, learn its workings, and iterate on their prompts to generate more refined and personalized images. The video concludes with a stunning example of an image generated through this process, highlighting the artistry and effort involved in creating beautiful and stable diffusion images.

Takeaways

  • 🎨 **Importance of Negative Prompts:** Binks emphasizes the significance of negative prompts in generating desired images with Stable Diffusion, avoiding unwanted elements.
  • 🖼️ **Avoiding Canvas Frames:** By including 'canvas frame' in the negative prompt, the AI is instructed to avoid generating images with frames.
  • 🚫 **Excluding Unwanted Features:** Negative prompts help prevent disfigurement, extra limbs, and mutations in the generated images.
  • 📝 **Crafting Tailored Prompts:** Binks demonstrates how to create detailed prompts to guide the AI towards the desired outcome.
  • 💡 **Sampling Method Recommendation:** DPM++ SDE Keras is Binks' preferred sampling method for protogen, often with 30 sampling steps.
  • 🔍 **Image Quality Enhancement:** Increasing the image height and config scale can improve the portrait look and quality.
  • 🧩 **Restoring Facial Features:** The 'restore faces' option is used to maintain facial integrity in the generated images.
  • 🎭 **Stylizing the Portrait:** A simple initial prompt like 'portrait of a blonde woman, medieval' sets a stylistic direction before applying negative prompts.
  • 📈 **Iterative Improvement:** Binks suggests using a basic prompt first and then refining it with more detailed negative prompts for better results.
  • 📚 **Learning the System:** There's an artistic aspect to working with AI, requiring understanding and experimentation to achieve the best results.
  • 🔄 **Persistence Pays Off:** Encouragement to repeatedly generate images and refine prompts to reach the desired outcome.

Q & A

  • What is the main topic of the video?

    -The main topic of the video is the importance of negative prompts in stable diffusion for generating images that match the user's expectations.

  • Who is the presenter of the video?

    -The presenter of the video is Binks.

  • What version of ProtoGen is Binks using in the video?

    -Binks is using ProtoGen version 3.4, specifically the photorealism release.

  • What is the significance of negative prompts in image generation?

    -Negative prompts are significant because they help the AI avoid generating unwanted elements in the image, such as canvas frames or disfigured subjects.

  • What sampling method does Binks prefer for ProtoGen?

    -Binks prefers the DPM++ SDE Keras sampling method for ProtoGen.

  • How many sampling steps does Binks usually set for image generation?

    -Binks usually sets the sampling steps to around 30.

  • What does Binks do to improve the portrait look of the generated image?

    -Binks increases the height from 512 to 768 to give the image more of a portrait look.

  • What option does Binks enable to improve the facial features in the generated image?

    -Binks enables the 'restore faces' option to improve the facial features.

  • How does Binks tailor the original prompt to get closer to the desired image?

    -Binks tailors the original prompt by adding more detailed descriptions and keywords to describe the subject in the desired style.

  • What does Binks suggest for users to do after watching the video?

    -Binks suggests that users should go out and repeatedly click the generate button, experimenting with different prompts and settings.

  • How does Binks describe the process of using AI for image generation?

    -Binks describes the process as involving a bit of work and artistry, where one must understand how the system works to achieve the desired results.

  • What does Binks encourage viewers to do if they have questions or find the video useful?

    -Binks encourages viewers to leave questions in the comment section and to like and subscribe if they find the video useful.

Outlines

00:00

🎨 Introduction to Negative Prompts in Stable Diffusion

In this introductory paragraph, Binks welcomes viewers back to a new video where they will discuss the significance of negative prompts in the context of stable diffusion and why generated images might not match expectations. Binks mentions using Protogen version 3.4, a photorealism release by Darkstorm 2150, and provides a link to it in the description. They also share a brief overview of the types of images that can be generated with this model and proceed to demonstrate the use of a negative prompt to refine the image generation process.

Mindmap

Keywords

Negative Prompts

Negative prompts are phrases or words that are intentionally included in the input to guide an AI image-generating system, such as Stable Diffusion, to avoid including certain elements in the generated image. In the video, Binks uses negative prompts to prevent unwanted features like a canvas frame or disfigured subjects in the generated portrait.

Stable Diffusion

Stable Diffusion is an AI model used for generating images from textual descriptions. It is a part of the broader field of generative AI and is known for its ability to create detailed and realistic images. In the context of the video, Stable Diffusion is the tool used to create the portrait of a blonde woman in a medieval style.

Photorealism

Photorealism is a style of art that attempts to create images that closely resemble photographs. In the video, Binks mentions using a 'photorealism release' of the Protogen model, which suggests that the generated images are intended to look as realistic as possible, akin to actual photographs.

Protogen

Protogen is a reference to a specific version or model used within the AI image-generating system. Binks uses Protogen version 3.4 in the video to demonstrate how to generate high-quality images. It is implied that Protogen is capable of producing a wide range of image types, including photorealistic ones.

Sampling Method

The sampling method refers to the technique used by the AI to generate the image based on the input prompts. DPM++ SDE Keras is mentioned as Binks' preferred sampling method for Protogen, which likely stands for Deep Probabilistic Model with Stochastic Differential Equations implemented in Keras, a popular machine learning library.

Sampling Steps

Sampling steps are iterations or stages in the image generation process that determine the level of detail and refinement in the final image. Binks increases the sampling steps to 30 for a more detailed image, which is a parameter that can be adjusted to improve the quality of the output.

Config Scale

Config scale is a parameter within the AI system that affects the intensity or strength of the image generation process. Binks sets the config scale to 10, which is suggested to work well with the specific model being used, indicating a balance between detail and computational efficiency.

Restore Faces

Restore Faces is an option or feature within the AI system that likely focuses on improving the quality or realism of faces in the generated images. Binks checks this option, suggesting that it is important for achieving a more human-like appearance in the portrait.

Model Shoot Style

Model shoot style refers to the aesthetic or presentation style of a photoshoot, often associated with professional fashion or portrait photography. Binks tailors the prompt to include 'model shoot style' to guide the AI towards creating an image that has a professional and stylized look.

AI Artistry

AI Artistry is the creative process of using AI tools to generate artistic content. Binks emphasizes the artistic aspect of working with AI, where understanding the system and crafting the right prompts are crucial for generating desired images, thus blending technology with creativity.

Tailored Prompt

A tailored prompt is a carefully crafted input designed to guide the AI towards generating a specific type of image. Binks provides an example of a tailored prompt that includes detailed descriptions and multiple keywords to refine the image generation process and achieve a particular style or outcome.

Highlights

The importance of negative prompts in stable diffusion for generating desired images.

Using ProtoGen version 3.4 for photorealism in image generation.

The option to explore other models on Darkstorm's Civis AI page.

Examples of image generation with the ProtoGen model are available on the page.

Utilizing a negative prompt to refine the image generation process.

The impact of the word 'portrait' on the inclusion of a canvas frame in the image.

Crafting negative prompts to avoid disfigurement and mutations in the generated image.

Choosing DPM++ SDE Keras as the preferred sampling method for ProtoGen.

Increasing sampling steps and image height for better portrait orientation.

Configuring the model scale to enhance the image generation process.

Restoring faces as a default option for more accurate human depiction.

The difference in image quality with and without the use of a negative prompt.

Tailoring the original prompt to achieve a more stylized medieval model shoot look.

Using detailed prompts and keywords to refine the image generation outcome.

The necessity of artistry and understanding the AI system for better results.

Encouragement to experiment with the generate button for different outcomes.

The value of community interaction through comments for tips and questions.

The enjoyment of working with AI that requires a bit of work and creativity.