Stable diffusion is a new AI model that can generate high-quality images from text prompts. However, it can sometimes generate images that are blurry, pixelated, or otherwise low quality.
Today I’m going to share with you a secret weapon that will take your image-creation skills to the next level: Stable Diffusion negative prompts. That’s right, negative prompts are a game-changer for image creation, and I’m going to show you why and how to use them.
What are Stable Diffusion Negative Prompts?
A negative prompt is a way to tell a text-to-image AI model what you don’t want to see in the generated image. It is a parameter that tells the model to avoid generating images that match the description of the negative prompt.
For example, you could use stable diffusion negative prompts to tell the model to avoid generating images that are blurry, pixelated, or have a certain color scheme.
In this blog article, we will discuss the basics of stable diffusion negative prompts. We will also provide a list of common negative prompts that you can use to improve the quality of your generated images.
Why Use Negative Prompts?
They can help you remove unwanted objects, styles, or abnormalities from the original generated image. For example, if you want to create an image of a woman you can use negative prompts like this: bad anatomy, disfigured features, blurry details ,cloned faces
Negative prompts can help you explore different variations and possibilities of your image. If you want to create an image of a cyberpunk forest by Salvador Dali, but you don’t want any trees or green color in it, you can use negative prompts.
Positive Prompt : cyberpunk forest by Salvador Dali
Negative Prompt : Green, Trees
If you’re trying to make a picture of a beach, but you want it to show only the shore without any people, you can achieve this by using negative prompts to take out those elements from the image. In other words, you can guide the process by telling the tool what not to include, which will help you get the picture you’re aiming for with just the beach and no human figures. This way, you can create a serene beach scene without any people in it.
Positive Prompt : Beach shore, beautiful masterpiece
Negative Prompt : Humans, crowd, people
Few other examples
Positive Prompt : Cute happy Dog, best quality, depth of field, ultra high res, intricate, photorealistic, masterpiece, ultra-detailed
Negative Prompt : semi-realistic, cgi, 3d, render, sketch, cartoon, drawing, anime, mutated hands antd fingers, deformed, distorted, disfigured, poorly drawn, bad anatomy, wrong anatomy, mutation, mutated, ugly, disgusting
Positive Prompt : Photo of a car, uhd, dslr, soft lighting, high quality, cinematic shot, soft lighting, detailed background
Negative Prompt : semi-realistic, cgi, 3d, render, sketch, cartoon, drawing, anime, mutated ,deformed, distorted, disfigured, poorly drawn, bad anatomy, ugly, disgusting, amputation
Here are some tips on using negative prompts for Stable Diffusion
- Use specific and descriptive words. Instead of using vague words like “bad” or “ugly,” use words that describe what makes something bad or ugly for you. For example, instead of saying “no bad images,” you could say “no blurry images,” “no distorted images,” or “no images with too many objects.”
- Use multiple words. You can use multiple words in a negative prompt, separated by commas. This will help to ensure that the model does not generate any images that contain any of the words in the negative prompt. For example, you could say “no blurry, distorted, or noisy images.”
- Use weights. You can use weights to adjust the importance of each word in a negative prompt. This can be helpful if you want to emphasize certain words or phrases. For example, you could say “no blurry images: -2.0, no distorted images: -1.0.” This would tell the model that it is more important to avoid blurry images than distorted images.
- Experiment. The best way to learn how to use negative prompts effectively is to experiment. Try different combinations of words, weights, and phrases to see how they affect the output of the model. You may be surprised at what you can achieve!
Stable Diffusion Negative Prompts
| Missing arms | Poorly drawn face | Text | Duplicate features | Long neck | Absent limbs |
| Malformed | Kitsch | Signature | Dehydrated | Unnatural angle | Poorly drawn face |
| Childish | Grainy | Extra legs | Poorly drawn feet | Unfocused | Multiple views |
| Incorrect ratio | Amputee | Macabre | Shortened | Tiling | Grains |
| Mutated hands | Morbid | Extra hands | Uncreative | Ugly | Mistake |
| Unattractive | Mistake | Distorted limbs | Blurry | Incorrect ratio | Bad eyes |
| Low quality | Incorrect physiology | Disconnected limb | Bad anatomy | Unsightly | Flaw |
| Unnatural pose | Revolting dimensions | Beyond the borders | Extra arms | Identifying mark | Lowres |
| Duplicate | Out of frame | Indistinct | Username | Unnatural pose | Pixelated |
| Disfigured | Poorly drawn hands | Watermark | Unsightly | Double face | Bad illustration |
| Missing fingers | Overexposed | Extra fingers | Low resolution | Deformed | Poorly drawn hands |
| Unfocused | Cloned face | Trimmed | Missing hands | Extra limbs | Extra arms |
Conclusion
Negative prompts are a powerful tool that can be used to improve the quality and consistency of the images generated by Stable Diffusion. By using negative prompts, you can tell the model what you don’t want to see in the generated images, which can help to prevent the model from generating blurry, distorted, or noisy images. You can also use negative prompts to modify the style of the generated images.
