Create your AI Cum Slut -70% for Mother's Day
x

AI [AI Art] - Show Us Your AI Skill *NO TEENS*

4.80 star(s) 6 Votes

devilkkw

Member
Mar 17, 2021
338
1,146
So you made image in sd1.5 and do i2i with flux?
What about your PC spec? Running flux on 6Gb is possible?
The flux1.dev model I use is 22.1GB and my graphics card is 24GB RTX3090... But Civitai has modified Flux models that also work with 6GB graphics cards.
Just for know, your method works with anime and toon model?
 

Sharinel

Active Member
Dec 23, 2018
613
2,586
So you made image in sd1.5 and do i2i with flux?
What about your PC spec? Running flux on 6Gb is possible?

Just for know, your method works with anime and toon model?
I'm doing this at the moment for faces, so that I can incorporate Loras then finish in Flux. Using the dev-fp8 version





I wasn't really going for anime or toon models but this one is semi-real and seems to have kept that 'style' when applying flux
 

sertsessiz

Newbie
Jul 27, 2017
65
732
So you made image in sd1.5 and do i2i with flux?
What about your PC spec? Running flux on 6Gb is possible?

Just for know, your method works with anime and toon model?
I haven't tried in anime and toon style visuals, but I know that XL or PONY models are already successful in such visuals. If they are used together with LORA suitable for the style...
 
  • Like
Reactions: devilkkw

devilkkw

Member
Mar 17, 2021
338
1,146
I'm doing this at the moment for faces, so that I can incorporate Loras then finish in Flux. Using the dev-fp8 version





I wasn't really going for anime or toon models but this one is semi-real and seems to have kept that 'style' when applying flux
interesting test. the face come more realistic, and rest of image are not touched. nice. thank you for take time and posting sample.
I haven't tried in anime and toon style visuals, but I know that XL or PONY models are already successful in such visuals. If they are used together with LORA suitable for the style...
yes, i know but i didn't like pony. all those "score_#" is not good, but what i see many user switched to pony. i keep working on sd1.5, but on some day i think i have to switch.
Also gave a chance to XL, but until i port my embeddings and lora to it, and do some merge, i'll keep use sd1.5.
Also i think CUI have pushed out the power of image generation, because you are allowed to experiment and have some parameters that aren't on a1111.
 
  • Like
Reactions: Sharinel

sertsessiz

Newbie
Jul 27, 2017
65
732
interesting test. the face come more realistic, and rest of image are not touched. nice. thank you for take time and posting sample.

yes, i know but i didn't like pony. all those "score_#" is not good, but what i see many user switched to pony. i keep working on sd1.5, but on some day i think i have to switch.
Also gave a chance to XL, but until i port my embeddings and lora to it, and do some merge, i'll keep use sd1.5.
Also i think CUI have pushed out the power of image generation, because you are allowed to experiment and have some parameters that aren't on a1111.
I totally agree, I stopped using A1111. I'm using SwarmUI, which has the ComfyUI interface in it. And it's easy to use... And then there's Forge... The interface is almost the same as A1111 but it's also compatible with Flux...
 
  • Like
Reactions: devilkkw
4.80 star(s) 6 Votes