[][][][]][] AI ART COMPETITION

[Stealthy_Enigma]

i need to know...why are you cranking the cfg this high.... please i need an answer

Cause otherwise the tail is ruined (shape, color, or amount), the eye color becomes either pure purple blobs, or not quite look right. I also tend to get stuff I don't like or want (human earlobes, random foxes appearing, actual fox head, etc.)

Spoiler: Examples

Most of my gripes are about the tail or face. I just mess around to minimize the defects, which usually ends up increasing the CFG to fix most of the issues. The way I see it, the higher the CFG, the closer it follows my prompt (for the most part).

I am no techno mage, so I just fiddle around until I get what I want, which sometimes majorly backfires on me (like the time I wanted to get stable diffusion on my computer, only to learn the hard way that I did not have enough ram to run it- after watching a lengthy tutorial and following all those confusing procedures properly).
˝<(✿;>⁠ᆺ<;⁠) >˝
/\ /˙˙˙˙˙⊃⊂\

 

[Alfir]

 

[greyblob]

Cause otherwise the tail is ruined (shape, color, or amount), the eye color becomes either pure purple blobs, or not quite look right. I also tend to get stuff I don't like or want (human earlobes, random foxes appearing, actual fox head, etc.)

Spoiler: Examples

View attachment 22794View attachment 22795View attachment 22796View attachment 22797

Most of my gripes are about the tail or face. I just mess around to minimize the defects, which usually ends up increasing the CFG to fix most of the issues. The way I see it, the higher the CFG, the closer it follows my prompt (for the most part).

I am no techno mage, so I just fiddle around until I get what I want, which sometimes majorly backfires on me (like the time I wanted to get stable diffusion on my computer, only to learn the hard way that I did not have enough ram to run it- after watching a lengthy tutorial and following all those confusing procedures properly).
˝<(✿;>⁠ᆺ<;⁠) >˝
/\ /˙˙˙˙˙⊃⊂\

negatives and prompt weights are used for exactly that.. instead of raising the cfg this high try adding (tail:1.5) and (extra tails:1.3) as negatives, for example.

also what's your vram?

 

[Stealthy_Enigma]

negatives and prompt weights are used for exactly that.. instead of raising the cfg this high try adding (tail:1.5) and (extra tails:1.3) as negatives, for example.

also what's your vram?

My computer has 8GB of ram wait-

Oh... I have 128 MB vram, probably nowhere near enough for what I need.
˝<(✿≖﹏≖;)>˝
/\/??\

 

[greyblob]

My computer has 8GB of ram wait-

Oh... I have 128 MB vram, probably nowhere near enough for what I need.
˝<(✿≖﹏≖;)>˝
/\˝/??\

yeeea.. that's not much. i mean you could still run it. it can even run on cpu. but it'll take a very very long time that'd just be tedious.

i think the lowest reasonable vram is 2gb.

 

[greyblob]

WINNER/S: @melchi @AliceShiki @Cipiteca396

 

[AliceShiki]

New theme... Fire?

 

[greyblob]

eighty first round:

-theme: fire
-model: Anything V5 on https://app.prodia.com/#/art-ai

you can play with anything else, cvg, sampler, etc.
optional: this site gives you full control on the prompt, so you might want to add these for some extra quality:

you can view the seed by uploading your image here: https://www.nayuki.io/page/png-file-chunk-inspector

positive: masterpiece, best quality, detailed

recommend negative by the model creator:
(worst quality:1.6, low quality:1.6), (zombie, sketch, interlocked fingers, comic)

nemu's negative:
lowres, ((bad anatomy)), ((bad hands)), text, missing finger, extra digits, fewer digits, blurry, ((mutated hands and fingers)), (poorly drawn face), ((mutation)), ((deformed face)), (ugly), ((bad proportions)), ((extra limbs)), extra face, (double head), (extra head), ((extra feet)), monster, logo, cropped, worst quality, low quality, normal quality, jpeg, humpbacked, long body, long neck, ((jpeg artifacts)), signature, watermark

 

[Cipiteca396]

eighty first round:

-theme: fire
-model: Anything V5 on https://app.prodia.com/#/art-ai

 

[greyblob]

View attachment 22822

ah i forget this was the horny model

 

[Cipiteca396]

ah i forget this was the horny model

Tbh, it's a portrait model. There are far hornier models.

 

[Stealthy_Enigma]

 

[AliceShiki]

Be careful! You're gonna burn your hat!!! >.<

On a side note, the prompt actually included Fireball, but this was the very first attempt and I liked it quite a bit, so I just went with it~

 

[Aaqil]

 

[greyblob]

WINNER: @AliceShiki

 

[greyblob]

does anyone have a model they'd like to continue using? I think anythingv5 has been the best so far for everyone

 

[melchi]

Maybe next prompt could be "Motion blur"

 

[greyblob]

eighty second round:

-theme: motion blur
-model: Anything V5 on https://app.prodia.com/#/art-ai

you can play with anything else, cvg, sampler, etc.
optional: this site gives you full control on the prompt, so you might want to add these for some extra quality:

you can view the seed by uploading your image here: https://www.nayuki.io/page/png-file-chunk-inspector

positive: masterpiece, best quality, detailed

recommend negative by the model creator:
(worst quality:1.6, low quality:1.6), (zombie, sketch, interlocked fingers, comic)

nemu's negative:
lowres, ((bad anatomy)), ((bad hands)), text, missing finger, extra digits, fewer digits, blurry, ((mutated hands and fingers)), (poorly drawn face), ((mutation)), ((deformed face)), (ugly), ((bad proportions)), ((extra limbs)), extra face, (double head), (extra head), ((extra feet)), monster, logo, cropped, worst quality, low quality, normal quality, jpeg, humpbacked, long body, long neck, ((jpeg artifacts)), signature, watermark

 

[melchi]

This one is a little harder to get good images with.

 

[AliceShiki]

does anyone have a model they'd like to continue using? I think anythingv5 has been the best so far for everyone

I did like it quite a bit! ^^)/