Image input and prompt explorations with stable diffusion img2img

Seeing what happens when you keep the input image constant and vary the prompts

Kawandeep Virdee
4 min readFeb 21, 2023

I was curious what would happen if I kept the same input image, and varied the prompt. This is following from the earlier exploration varying the strength with the same input image. You can see that here:

Here’s the input image I’m starting with:

Now here’s what happens varying the prompts! The top of each image includes the strength and the prompt, and I’ve included the prompts in the captions too. These are just simple one to two word prompts.

Cheese, Forest, Buildings
sandwich, swimming pool, space
sun, a beautiful home, a fantasy landscape
a creature, a monster

Now here’s the same input drawing from the earlier post, with different prompts. Input image:

Here are the generated images, varying the prompt:

Vegetables, spaceships, castles
castles, super mario, the matrix
rainforest, clouds, a sitting yogi meditating
cornfields, at strenght 0.8 and 0.7
a cozy interior, forest
space disco, fruits

Thanks for reading! You can follow the latest of what I’m up to on medium and on twitter. You can get monthly recaps of what I’m up to on this mailing list. I’m interested in starting conversations, hearing ideas, and finding opportunities related to what I’m working on. If you know folks who would be into this, I’d love it if you share this with them! And if you’d like to jam more on any of this, I’d love to chat here or on twitter.



Kawandeep Virdee

Building. Author of “Feeling Great About My Butt.” Previously: Creators @Medium, Product @embedly, Research @NECSI.