JENNIE For BLACKPINK - CivArchive (CivitAI Archive)

JENNIE For BLACKPINK - V1.0

Description

FAQ

Comments (12)

TaihouMar 7, 2023· 2 reactions

CivitAI

"Girls frontline" tag? What?

MaxfieldMar 19, 2023

CivitAI

only generate images are allowed.

No QR codes are allowed in your images. You have been muted and prevented from commenting, reviewing, or uploading for 48 hours as a warning.

maxhitmanApr 4, 2023

CivitAI

What is the "Trigger Word" please ?

shangrila1985437Apr 6, 2023

CivitAI

what check point for this?

AIGeishaApr 8, 2023

CivitAI

very low quality. do not waste time.

ClementinaApr 21, 2023· 3 reactions

CivitAI

Strong disagree that this is "very low quality". This is an extremely strong LORA, but highly trained so you'll need to reduce its weight to 0.3-0.6 to get the best results.

GSwanMay 18, 2023

Did you try using something else than ChilloutMix? I am not able to get a good render even with weight 1 - 1.5

ClementinaMay 18, 2023· 1 reaction

@GSwan You want weight 0.4-0.6. I've used this LORA successfully with a wide variety of models -- I'll post some examples below.

GSwanMay 18, 2023

CivitAI

which one is the trigger word?

ericdawnmerJun 3, 2023

CivitAI

Please do Mai Shiraishi (白石麻衣) or Erika Ikuta (生田絵梨花) or Nanase Nishino (西野七瀬)😄

haofengJun 25, 2023

CivitAI

I always generate multi-people when I set resolution 1024*2048. I use negative prompts like "multi-people" and "duplicate", but it not work, anyone has ideas?

SivviNov 9, 2024

Most SD1.5 models are trained with a base resolution of 512 pixels in width. If you set the canvas size to 1024x2048px, the AI model interprets it as having two or more canvases to render the image, therefore the duplicate / multi people result.

To generate a large image, start with a 512x1024px resolution, then move the image to the img2img tab. Use the same prompt, adjust the denoise strength (ideally between 0.3 and 0.6), and set resize by scale 2x.

LORA

SD 1.5