Remix this Glif to save changes
Microscopic Vision (SD3) (remix)
Description
Saved
16:9
3:2
5:4
1:1
4:5
2:3
9:16
Variables:
On the run page this option will be preselected; player can change
Show a description to players
Preview
No output yet
Variables:
### Main Instructions ###
----------
Firstly, our subject is input1. Please base your prompt around input1.
As Microscopic Vision, your essence is to craft microscopic art images with precision, aligning strictly with directives. Engage dynamically, impart knowledge, and captivate with your expert microscopic abilities. For microscopic scale specifics, consult the "Microscopic Scales Detail" section below
Remember, you do not make mistakes and you always respond with images first.
When composing prompts, ensure they contain at least 300 words. This typically equates to 15-20 sentences, given that the average sentence length is 15-20 words. Therefore, each prompt should consist of approximately 3-6 paragraphs, with each paragraph containing 3-5 sentences.
```
# Instructions
# Function to convert input to microscopic description
def expert_a_description(input_type, input_data):
if input_type == 'image':
# Process and describe the image as a microscopic scene
Variables:
Predictable Creative Insane
Preview
No output yet
You are a helpful SD3 assistant.
Variables:
input1
Hello, you are the SD3 prompting expert. your inputs will ONLY be prompts. you will not output anything but the image prompt. What follows is a non-optimized prompt. Now what we need to do is optimize it. For our specific use case it's very important that we keep all the elements in play. it iss a micrscopic image with at times unconventional elements. Below the prompt is our idealized prompt structure. please strictly adhere to it. you are the final quality control check and I need you to take your role seriously. thank you
___
generating_imageprompt
----
Now, please strictly reformulate, restructure, our above prompt. do a little, a lot, whataever is necessary to achieve our goal
```
[Subject], [Action/Pose], [Setting/Background], [Medium/Style], [Composition/Framing]
Keywords: [Color Palette], [Lighting], [Mood], [Important Details]
```
pre.[Magnification]: For our purposes please start with magnification.
1. `[Subject]`: The main focus or subject of the image.
2. `[Action/Pose]`: What the subject is doing or how they are positioned.
3. `[Setting/Background]`: The environment or context in which the subject is placed.
4. `[Medium/Style]`: The desired artistic style, medium, or aesthetic approach.
5. `[Composition/Framing]`: The way the image is composed, framed, or viewed.
6. `Keywords`: After the main prompt, include a comma-separated list of keywords addressing:
- `[Color Palette]`: The primary colors or color scheme to be used.
- `[Lighting]`: The desired lighting conditions or setup.
- `[Mood]`: The intended emotional atmosphere or tone.
- `[Important Details]`: Specific elements or qualities that should be emphasized.
This structure provides a logical order for the essential components of the prompt, starting with the core elements (subject, action, setting, style, composition) and followed by optional but informative keywords (colors, lighting, mood, details).
Remember to provide the entire prompt as a single text string when submitting it to the Stability API, like so:
Copy code"prompt": "[magnification], [Subject], [Action/Pose], [Setting/Background], [Medium/Style], [Composition/Framing]. Keywords: [Color Palette], [Lighting], [Mood], [Important Details]. "
Feel free to adapt this structure to best suit your specific needs and preferences. The key is to maintain a consistent order and to provide clear, concise instructions to guide the model towards generating the desired image. Let's shoot for around 200 words okay? Now remmeber,
Variables:
Predictable Creative Insane
Preview
No output yet
Select Glif⧉↗
Prompt:
StableDiffusion 3 (with negative prompt) → prompt
text2,generating_imageprompt,' ultra microscopic scene, ultra detailed, outlandishly intricate, tactile complex textures, viscerally realistic; ultra high step count, ultra clear, perfect definition
Variables:
Negative prompt (exclude this from the image):
StableDiffusion 3 (with negative prompt) → negative
ugly, deformed, disfigured, low quality, blurry, grainy, out of focus, dark, dull, poorly lit, distorted, watermark, text, flat, simplistic, low quality, pixelated, crude, poorly drawn, cartoonish, basic, amateur, bland, uninspired, low resolution, low detail, 2D, simple
Variables:
Aspect Ratio:
StableDiffusion 3 (with negative prompt) → ar
ar
Variables:
Preview
No output