Damn man, I'm sure you've already been asked before (and my apologies if you don't like that), but are you perhaps willing to explain your creative process for these images? Or at the very least just point me in the right direction of what software/guides I should be looking at?
The quality is seriously impressive!
curlyjoe
Member
- Dec 18, 2022
- 411
- 3,354
- Sep 20, 2021
- 1,097
- 5,871
- Jan 6, 2021
- 6,463
- 58,163
- Jun 8, 2018
- 833
- 5,625
Sure. I can explain in private more so that we don't turn this thread into a long description of my workflow, but basically I'm using ComfyUI. This is the web interface that allows me to select models, loras, samplers, prepare the prompt and adjust the image wherever it's needed.
In ComfyUI you can create multiple workflows. In my case I have one workflow that generates random girls based on some lists I have made (body parts, face features, hair styles and so on) and then I generate maybe 10-20 girls out of which I select one which I copy to the posing workflow where I am composing a prompt in multiple nodes. For consistency reason my prompt building tool takes 5 nodes: quality, pose, character, clothing and environment.
Usually I have to generate 10-20 images for each pose I want with a character so that I can select one that matches the rest.
If you're not familiar with ComfyUI there are a lot of tutorials on youtube. It can be overwhelming at first but after some time you'll get really used to it.
