【GK】Functional: More types of camisole variations丨PONY丨SDXL

0
0
0
0
RealisticPortrait PhotographyGirlElement Boost
Actualizado recientemente: Publicado por primera vez:
Realistic,Portrait Photography,Girl,LoRA,PonyImage info
Realistic,Portrait Photography,Girl,LoRA,PonyImage info
Realistic,Portrait Photography,Girl,LoRA,PonyImage info
Realistic,Portrait Photography,Girl,LoRA,PonyImage info
Realistic,Portrait Photography,Girl,LoRA,PonyImage info
Realistic,Portrait Photography,Girl,LoRA,PonyImage info

[Camisole variant v1.0 version]

Remember to return the pictures!

Its characteristics are the styles of variant suspenders & vests and even tube tops, the realistic effect is good, the color can be changed, the 7th level avatar has the best effect, and it also has a whitening effect! (It will slightly change the picture to a certain extent)

The sample pictures are all produced locally in 4090, and the effects are different from those generated online!

Recommended large models: PONY large model waiREALCN series, (other branches have not been tested yet, you are welcome to try different large models and return pictures)

<prompt word>

Recommended LoRa weight: 0.8~1.2

Trigger words: Expressiveh,shoulders

Assist words (add more elements): woman focus, concept art, shiny_skin, garter belt, tallhighs (it is recommended to add some skin texture and leg detail decoration, as well as prompt words for clothing details to control the composition output effect, such as lace edges , ribbon, etc.)

Recommended prompt words: score_9, score_8_up, score_7_up, score_6_up, score_5_up, score_4_up, ultra detailed, perfect face, HD32K, intricate

Recommended negative prompt words: score_6, score_5, score_4,

The PONY model itself does not require too many quality modifiers

Everyone is welcome to add and try using different elements or commonly used prompt words~



<Generate parameters>

Sampling method:Eulera

Number of iteration steps: 25+

Recommended sizes: 1:1/2:3/3:2/3:4/4:3/9:16/16:9 (other sizes are welcome to try)

Prompt word guidance coefficient (CFG):5-7

It is recommended to enable AD direct output at 800*1200 resolution, or perform 1.1~1.5x high-definition repair

If you are pursuing higher precision and use high-definition amplification, the redrawing range is recommended to be 0.35-0.5.

Recommended for HD enlarged models: 4x-UltraSharp



<Notes>

There is a chance that problems with the hands, body, and limbs will occur during the generation, and multiple card draws will be required!

If conditions permit, it is recommended to download it for local use~

The local model name is for archiving, so you don’t need to mind the LORA model name in the example diagram, just use it normally.

It is recommended that the output resolution should not be lower than 800*1200.

When the resolution is low, the character's face may appear collapsed and blurred.

It is recommended to open ADetailer and check Enable, and select face yolov8n to repair!

Discusión

Más popular
|
Más nuevo
Enviar
Próximamente
Descargar
(0.00KB)
Detalles
Tipo
Recuento de generación en línea
0
Descargas
0
Parámetros Recomendados
Combinación Recomendada de Checkpoint
Ninguno
Peso recomendado
1
CFG
7
VAE
Ninguno

Galería

Más popular
|
Más nuevo