Stable Diffusion Negative Prompt Tips

Today, we will discuss the topic of Stable Diffusion Negative prompt. We will cover what negative prompt is and provide tips on how to use it more effectively and efficiently. 😀😀

What is Negative prompt 

This powerful method helps to refine the output of SD models by excluding unwanted elements so that you can control the output in a more accurate way.

How to use Negative prompt

In the process of generating images with Stable Diffusion Web UI, you can just simply input the elements (Negative Prompts) that you want to filter out in your images.

I list some practical negative prompts below:

👍Universal Negative Prompts

weird, worst quality, normal quality, low quality, low res, blurry, text, watermark, logo, banner, extra digits, cropped, signature, username, error, duplicate, ugly, horror, mutation, disgusting

🦄Negative prompts for an anime character

bad anatomy, bad hands, three hands, three legs, bad arms, missing legs, missing arms, poorly drawn face, bad face, fused face, cloned face, worst face, three crus, extra crus, fused crus, worst feet, three feet, fused feet, fused thigh, three thigh, fused thigh, extra thigh, worst thigh, missing fingers, extra fingers, ugly fingers, long fingers, horn, realistic photo, extra eyes, huge eyes, 2girl, amputation, disconnected limbs

😉Negative prompts for realistic character

bad anatomy, bad hands, three hands, three legs, bad arms, missing legs, missing arms, poorly drawn face, bad face, fused face, cloned face, worst face, three crus, extra crus, fused crus, worst feet, three feet, fused feet, fused thigh, three thigh, fused thigh, extra thigh, worst thigh, missing fingers, extra fingers, ugly fingers, long fingers, horn, extra eyes, huge eyes, 2girl, amputation, disconnected limbs, cartoon, cg, 3d, unreal, animate

🏘️Negative prompts for street views

overly empty and deserted, apocalyptic, overly crowded, cluttered, floating objects, distorted perspective, blurry cars/objects, strangely parked cars, vehicles cut off, signs you can't read

⛰️Negative prompts for landscapes

oversaturated, people, cluttered, busy, crowded, signage, animal, cars, plane

🍔Negative prompts for pictures of food

utensils, napkins, hands, arms, distracting backgrounds, non-food items, poor lighting, overexposed, underexposed, blurry, unfocused

🔞Negative prompts for non-adult content

nsfw, censored, nude

How to simplify the use of Negative prompt

Stable Diffusion embeddings, also known as textual inversion embeddings, are a method used to define new keywords in a model without modifying it. 

These embeddings are low-dimensional vectors that map features of an object or style and guide the Stable Diffusion model to produce images that match the user's input. It can be utilized for a series of relevant negative prompts.

The most widely used negative prompts embedding

EasyNegative addresses many fundamental issues and is very simple to use.

Designed to address hand-related problems

With badhandv4, you can improve the hand details of AI-generated images with less impact on the style of painting. 

Place the downloaded file, badhandv4.pt, in the 'embeddings' folder of your stable diffusion directory. After that, you just need to enter 'badhandv4' in the negative prompt field to make it effective.

Negative Embedding for Realistic Vision v2.0

The Negative embedding for Realistic Vision is intended to be used with this model to enhance the visual appeal of generated characters.

Deep Negative

Deep Negative can enhance the composition and colors of images, reducing instances of distorted facial features, incorrect body structures, and inverted spatial arrangements. It is suitable for large models, whether in anime or realistic styles.

Fast Negative

Fast Negative is also a powerful negative keyword embedding model. It incorporates commonly used negative keywords and can enhance image details with minimal impact on the original art style. It is suitable for large models in both anime and realistic styles.

Prompt
masterpiece,1girl, 25,upper body, looking at viewer, sitting at the edge of a river, in the forest

Negative Prompt
| realisticvision-negative-embedding

Parameters
Steps: 20,
Sampler: DPM++ 2M Karras,
CFG scale: 7,
Seed: 2650588257,
Size: 576x768,
Model hash: ab5d1ecdb2,
Model: Realistic Vision V4.0_V4.0,
Clip skip: 2,

You can definitely see significant differences between these two images.🥰