Create your own fiery horse images with our free AI image generator. It's easy and fun!


Join the conversation
If you strive to have SD make pictures that reflect exactly what you're visualizing, adjust your expectation because SD is best when given it's head. Just because the result doesn't match your expectation isn't a failure on SDs part, it's a 'failure' of imagination. Seize on what you get and let those unanticipated results take you for a walk... Overthinking (doing) the Promt: The 'Beautiful majestic Chinese woman is an example of this. I got very usable results in 2.1 with just the First Prompt and a generic Negative one. From what I've seen of your prompts, they check the boxes - in fact they look like Prompts from the old Stability.ai Prompt book except that they lack 2.1 specific 'magic words.' continued
If you strive to have SD make pictures that reflect exactly what you're visualizing, adjust your expectation because SD is best when given it's head. Just because the result doesn't match your expectation isn't a failure on SDs part, it's a 'failure' of imagination. Seize on what you get and let those unanticipated results take you for a 'walk.' Using only the first Prompt and a generic Negative Prompt, I got some cool results right of the bat so one thing I would say is that you are, perhaps, overdoing the Prompt Continued...
Ok... First let me say you already have excellent chops and great taste. Looked into 3 different Creations in detail: The 'WLOP' ad infinitum Prompt did exactly what I expected in both 1.5 and 2.1 - 2.1 was a near total flop whereas 1.5 returned some usable images. The Colors Piece returned nice images in 2.1 the first time I ran it (OH! Super Kudos for running with OPEN PROMPTS!) Btw. Greg Rutkowski wishes people wouldn't invoke his 'style' if you give a damn about that. There are SO many more Artists to choose from - study what's out there and try them. I once spent a day (and lots of credits) simply swapping-in different artists and combinations. continued...
Hi... please can you give me some tips on stable diffusion... I have to evolve my images more than 1000 timea to get an average results... (If possible, look at few of my latest creation) thanks ☺️
Low
1:1
Medium
90%
K_DPM_2_ANCESTRAL
SIMPLE

like these: detailed skin texture ultrarealistic uhd faces uhdi trending on pexels diffused natural skin glow breath-taking and beautiful subsurface scattering, striking hyper real painting Try adding some of these and see what happens: Here's an example of what I'm getting:
/jobs/aCOGJZVqdHPtroRz40W8/aCOGJZVqdHPtroRz40W8--1--4zhwm.jpg?tr=w-1600,c-at_max