FluxAIGridComparisons
A collection of various image grids created with Flux. This repository showcases comparisons of different prompts and their results, including variations in hair styles, clothing, nationalities, ages, and more.
About
This project aims to demonstrate the capabilities and versatility of Flux AI in generating diverse image sets. By comparing different prompts and their outputs, we can explore how subtle changes in prompts affect the generated images.
Examples
Here are some example prompts and their metadata used in this project:
1. Man Chopping Wood
Prompt:
all_metadata["amanchoppingwoo/123998"] = {
"sui_image_params": {
"prompt": "A man chopping wood axe raised above his head",
"model": "Flux/flux1-dev-fp8",
"seed": 123998,
"steps": 25,
"cfgscale": 0.8,
"aspectratio": "Custom",
"width": 512,
"height": 768,
"sampler": "euler",
"scheduler": "simple",
"nopreviews": true,
"wildcardseed": -1,
"imageformat": "JPG",
"vae": "ae.sft",
"negativeprompt": "",
"swarm_version": "0.9.1.1",
"date": "2024-08-09",
"generation_time": "240.52 (prep) and 59.92 (gen) seconds"
}
}
2. Man Climbing Rock Wall
Prompt:
all_metadata["amanclimbingaro/123998"] = {
"sui_image_params": {
"prompt": "A man climbing a rock wall reaching for the next hold",
"model": "Flux/flux1-dev-fp8",
"seed": 123998,
"steps": 25,
"cfgscale": 0.8,
"aspectratio": "Custom",
"width": 512,
"height": 768,
"sampler": "euler",
"scheduler": "simple",
"nopreviews": true,
"wildcardseed": -1,
"imageformat": "JPG",
"vae": "ae.sft",
"negativeprompt": "",
"swarm_version": "0.9.1.1",
"date": "2024-08-09",
"generation_time": "16813.83 (prep) and 58.69 (gen) seconds"
}
}
3. Man Climbing Rope
Prompt:
all_metadata["amanclimbingaro38/123998"] = {
"sui_image_params": {
"prompt": "A man climbing a rope legs wrapped around it",
"model": "Flux/flux1-dev-fp8",
"seed": 123998,
"steps": 25,
"cfgscale": 0.8,
"aspectratio": "Custom",
"width": 512,
"height": 768,
"sampler": "euler",
"scheduler": "simple",
"nopreviews": true,
"wildcardseed": -1,
"imageformat": "JPG",
"vae": "ae.sft",
"negativeprompt": "",
"swarm_version": "0.9.1.1",
"date": "2024-08-09",
"generation_time": "119.41 (prep) and 58.11 (gen) seconds"
}
}
Structure
The repository is organized into different categories, each containing image grids and their corresponding prompt information. Some of the categories include:
- Hair Styles
- Clothing Variations
- Nationalities
- Age Groups
- Occupations
- Environments
- etc..
Each category folder contains the generated image grids and a JSON file with the prompts and metadata used to create them.
Usage
The prompt information provided can help you understand how different parameters and prompt structures affect the generated images.
Acknowledgments
Special thanks to the Flux AI team for creating such a powerful and versatile image generation model.
Top comments (0)