Spaces:

---

stabilityai

/

stable-diffusion

like 11.2k

Running on CPU Upgrade

Great results!

Also, I noticed your other reliable prompts that consistently produce good results (cyberpunk city, yul brynner as professor x) seemed to roughly follow the pattern of:

<1st phrase describing scene>, <2nd phrase describing scene>, <list of non-scene-specific adjectives (e.g. "intricate", "octane render", "photo-realistic", etc)>

Examples are "A cyberpunk bladerunner styled city, a marketplace full of people, ...", "borderlands gunslinger, head and shoulders portrait, ...", "portrait of yul brynner as professor x, character portrait, ..." (with the exception of prepending emphasised words like "symmetry!!" to the front). Some other reliable prompts I've seen also seemed to roughly follow this pattern.

I wonder if this "one-two punch" of 2 scene phrases followed by generic adjectives is actually a reliable template for making consistent prompts, or if it's just a coincidence.

P.S. there's actually a minor typo in your borderlands prompt: you wrote "octane tender" instead of "render" 🍗😂 (although I wonder if minor spelling mistakes like that actually make a noticeable difference, or if the AI somehow recognises that "octane tender" is close to "octane render" and generates a similar effect 🤔)

Thanks @SaeedB . I am still learning and continue to play around with the prompt composition. Practice makes perfect.

I have watched numerous YouTube videos on Stable Diffusion and others on prompt creation, and also take inspiration from other creators on Lexica, KREA, and other sites,... but usually try to structure my prompts in ways that have worked for me in the past.

If interested, here is a reference where I have saved some of my AI Art (and prompts in a lot of cases)... to make sure I have a record for future use 😀.
https://www.pinterest.com/crescent1967/my-created-ai-artwork/

Enjoy and good luck on your art creation.