How to generate illustration images using AI (Stable Diffusion)

Thu, 06 Apr 2023 00:43:19 +0900

What is Stable Diffusion?

Stable diffusion is an AI developed by a research team at the University of Munich in Germany that generates images from inputted text information. By training on various images, it is possible to generate a wide variety of images, from photorealistic ones to illustrations.

This time, I will introduce how to generate illustration images using the pre-trained data of Stable diffusion.

What you need

Google Account

That’s all.

Generation steps

Open https://colab.research.google.com
Select File on the top left and choose New notebook
Select Edit and choose Notebook settings
Change the Hardware accelerator to GPU
Paste and execute the following code

1

!pip install diffusers==0.8.0 transformers

Paste and execute the following code

1

from diffusers import StableDiffusionPipeline

Paste and execute the following code

1
2


pipe = StableDiffusionPipeline.from_pretrained("gsdf/Counterfeit-V2.5")
pipe.to("cuda")

Paste and execute the following code

1
2
3
4
5


prompt = "((masterpiece,best quality)),1girl, solo, animal ears, rabbit, barefoot, knees up, dress, sitting, rabbit ears, short sleeves, looking at viewer, grass, short hair, smile, white hair, puffy sleeves, outdoors, puffy short sleeves, bangs, on ground, full body, animal, white dress, sunlight, brown eyes, dappled sunlight, day, depth of field"
n_prompt = "EasyNegative, extra fingers,fewer fingers"

image = pipe(prompt, negative_prompt = n_prompt).images[0]
image

The Prompt used here is based on the Prompt at https://huggingface.co/gsdf/Counterfeit-V2.5.

Image Generation on kenji.blog

How to generate illustration images using AI (Stable Diffusion)

What is Stable Diffusion?

What you need

Generation steps

Generation results (some)

References