Open WheatTail opened 5 months ago
i have same issue, sometime it works that give me right number of character but most of time it just one
I'm using lastest forge, mcfgPDXL_v10 model, proper SDXL vae and this extension. I'm trying to figure out how this even work because it's doesn't match with readme.
For example: I have "base prompt" and "common prompt" enabled. Divide ratio is
1,1,1
. My prompt is:score_9, score_8_up, score_7_up, score_6_up, score_5_up, score_4_up, realistic, outstanding, detailed beautiful eyes, countershading, detailed soft lighting, ear fluff, cinematic vintage photography, sitting on couch BREAK night time, full moon, 3girls, (night city) behind the window, night, safe BREAK Twilight Sparkle, semi-anthro, thoughtful, frowned BREAK Fluttershy, semi-anthro, shy, timid look BREAK Rarity, semi-anthro, alluring, smile, teasing
As
score_9, score_8_up, score_7_up, score_6_up, score_5_up, score_4_up, realistic, outstanding, detailed beautiful eyes, countershading, detailed soft lighting, ear fluff, cinematic vintage photography, sitting on couch
is common prompt - it should be added to each character prompt.night time, full moon, 3girls, (night city) behind the window, night, safe
should be the base of image.Sounds pretty simple but in fact I got anything but what I want. Deformed characters, fused in one, wrong number of characters, all that in the same picture sometimes.
Even if I disable both common and base prompts and write everything manually:
score_9, score_8_up, score_7_up, score_6_up, score_5_up, score_4_up, realistic, outstanding, detailed beautiful eyes, countershading, detailed soft lighting, ear fluff, cinematic vintage photography, sitting on couch, safe, Twilight Sparkle, semi-anthro, thoughtful BREAK score_9, score_8_up, score_7_up, score_6_up, score_5_up, score_4_up, realistic, outstanding, detailed beautiful eyes, countershading, detailed soft lighting, ear fluff, cinematic vintage photography, sitting on couch, safe Fluttershy, semi-anthro, shy, timid look BREAK score_9, score_8_up, score_7_up, score_6_up, score_5_up, score_4_up, realistic, outstanding, detailed beautiful eyes, countershading, detailed soft lighting, ear fluff, cinematic vintage photography, sitting on couch, Rarity, safe, semi-anthro, alluring, smile, teasing
I'll got something like this:
I just can't get how to make it work properly. I've tried numerous times with and without common/base prompts are enabled. Different image sizes. But 99% times I got deformed/fused characters. Sometimes it give me almost what I want but as far as prompt become slightly more complicated it all just breaks.
score_9, score_8_up, score_7_up, score_6_up, score_5_up, score_4_up, Twilight Sparkle, semi-anthro, safe BREAK score_9, score_8_up, score_7_up, score_6_up, score_5_up, score_4_up, Fluttershy, semi-anthro, safe BREAK score_9, score_8_up, score_7_up, score_6_up, score_5_up, score_4_up, Rarity, semi-anthro, safe
score_9, score_8_up, score_7_up, score_6_up, score_5_up, score_4_up, Twilight Sparkle, semi-anthro, safe, magic, glowing horn BREAK score_9, score_8_up, score_7_up, score_6_up, score_5_up, score_4_up, Fluttershy, semi-anthro, safe, flying BREAK score_9, score_8_up, score_7_up, score_6_up, score_5_up, score_4_up, Rarity, semi-anthro, safe, sitting
idk if you saw the guide on civitai but I followed this: https://civitai.com/models/339604/how-to-generate-multiple-different-characters-mix-characters-andor-minimize-color-contamination-or-regional-prompt-adetailer-and-inpaint-or-my-workflow
ALSO, Do not use Base Prompt, it messes up the results, just use Common Prompt, it's a huge difference. and be sure to leave the Lora settings at 0 as well.
yeah dosent work half the time
Same here
I'm using lastest forge, mcfgPDXL_v10 model, proper SDXL vae and this extension. I'm trying to figure out how this even work because it's doesn't match with readme.
For example: I have "base prompt" and "common prompt" enabled. Divide ratio is
1,1,1
. My prompt is:As
score_9, score_8_up, score_7_up, score_6_up, score_5_up, score_4_up, realistic, outstanding, detailed beautiful eyes, countershading, detailed soft lighting, ear fluff, cinematic vintage photography, sitting on couch
is common prompt - it should be added to each character prompt.night time, full moon, 3girls, (night city) behind the window, night, safe
should be the base of image.Sounds pretty simple but in fact I got anything but what I want. Deformed characters, fused in one, wrong number of characters, all that in the same picture sometimes.
Even if I disable both common and base prompts and write everything manually:
I'll got something like this:
I just can't get how to make it work properly. I've tried numerous times with and without common/base prompts are enabled. Different image sizes. But 99% times I got deformed/fused characters. Sometimes it give me almost what I want but as far as prompt become slightly more complicated it all just breaks.