Stylized Blend of AutismMix SDXL (pony). Same prompting as ponydiffusionXL.
General prompting:
Positive: score_9, score_8_up, score_7_up, score_6_up, score_5_up, score_4_up, [describe the image], tag1, tag2...
negative:
For more complex compositions use tags from danbooru.
Personally using:
Euler a 30 steps
ADetailer with just face_yolov8n.pt with no prompts
1024x1024 or 832x1216/1216x832 (tall/wide) or 886x1182/1182x886 (photo) resolutions
Description
FAQ
Comments (13)
Looks good. How to intall this FreeU that you linked?
Copy that link into the "URL for extension's git repository" in Extensions>Install from URL in A1111 and click install. Then reload UI and FreeU is somewhere under ADetailer in txt2img or img2img tabs. Gotta select the drop down for SDXL recommendations in FreeU then click the ✅.
@MidFR i installed it but dont see any freeu tabs in a1111
i found other link on reddit and it worked, anyway thanks! https://github.com/ljleb/sd-webui-freeu
@Rating_Agent must have linked the wrong one mb
heya, will try this out!
if score_9, score_8_up, score_7_up still apply, does source_anime, source_furry, source_pony, do those still work? should they be used?
clip skip 2 needed still?
Needs clip skip 2. The source tags would still work but I didn't need them when I was testing.
when i make standing characters, they kind of show up in a base like if they were minis, how do i stop that? (posted photo to showcase)
No clue. Would be an issue from ponyXL I would ask there.
Really liking this model generally, nice style. However, I noticed that negative prompts can really change the style, making lines thinner and making the colours less defined etc, do you know how I might counteract this? I've linked a comparison, first without negative prompt and second with.
Looks like this gives great results.
So why isn't this available for creation on CivitAI?
Details
Available On (2 platforms)
Same model published on other platforms. May have additional downloads or version variants.