Detalhes
Recomendado
v1.5
GDXN mixed furry model

GDXN mixed furry model

8.9K
168
9
#Anime
#Personagem
#kemono
#Onde, cara?
#Peludo
#Desenho animado
#Anime
#Personagem
#kemono
#Peludo
#Desenho animado
#1girl
#Peludo

I got an idea of mixing some art styles lora of furry several days ago. So I created this model. The grapefruit v4.1 model by lkena is its basic model, and this is grapefruit model's homepage https://civitai.com/models/24383/grapefruit-hentai-model.

I mixed three furry art style lora model which were created by myself in it and the results were not bad. The three mixed lora models' original artists are DAGASI, shiitakemeshi and nefuraito336 on e621. They could be find on my civitai page.

By using this furry model

you could easily create:

1.furry/kemono female safe pictures

2.furry/kemono female ???? pictures

3.realistic furry/animal pictures

4.anime/realistic human girl's safe/nsfw pictures

not easy to create:

1.horizontal pictures

2.male creatures

3.multiple characters

4.correct hand anatomy

5.pictures with overly bright colour

Because the model is based on grepefruit ?????? model which is focus on female, you should add lots of tags to pledge the validity of male's ???????????? if you want to create some male's pictures. you could copy my prompts of the example pictures, and the nagetive prompts I used just the easiest NovelAI's nagetive prompts. But the model may generates some pictures with wrong creature ??????? because furry character's particularity, you could add some tags to correct it. I use the VAE "vae-ft-mse-840000-ema-pruned" to create the pictures. I haven't try the effect of Hire.fix on this model.

clip recommendation:

  • clip1: the worst style which is near the furry character created by Novel AI model.

  • clip2: more better than clip1, easy to create the pictures like the example pictures, but do not good at some details of eyes and the things with overly shiny colour.

  • clip3: the art style is near to clip2, the best eyes' details, but the effect of light and body fur details are little lower than clip2.

  • clip4/clip5: the difference between clip4 and clip5 is small. Although their art styles are not so bad, they are both not very good at following the tags' request.

  • clip6: a different art style which is a bit detailed than clip2/clip3, looks good as clip2/clip3, but I haven't do more test of it.

tips of tag-writing

the furry character: furry, furry female,

the not furry character: anthro, human, humanoid

the character's body color:

1.for the furry character: pale fur, pale body fur, beige fur, beige body fur and so on.

2.for the not furry character: pale body, beige body, blue body and so on.

the fur of character's body: smooth fur, detailed body fur, shiny fur or something

  • total tips: enter the character's species what you want to create, apart the tags which include words of colour by adding several detailed tags so that you could get the character's correct fur colour easily. If there are some mistakes of ??????? such as wrong tail number, bad hands or something, you could add some tags in prompts like "correct tail place", "correct tail ???????", "correct hand ???????" and so on. Besides, this model could create some humanoid but the effect is not good as creating furry. Avoid describing the color of the not furry character's skin by using tags like "pink skin". On the contrary, tags like "pink body" are more better for not furry character.

disclaimer

All my models are only for communication. Please don't use my models to create illegal pictures and the responsbility for generating relevant pictures will be brone by the pictures' prompt operator himself. What's more, if my example pictures violated your copyright, please contact me and I will delete my model and give an apology to you.

Ver tradução

Classificação e Avaliação

5.0 /5
0 classificação(ções)

Sem classificações ou avaliações suficientes

L
Anúncio
2024-04-02
Publicar Modelo
2023-05-06
Atualizar info do modelo
Detalhes do modelo
Tipo
Checkpoint
Tempo de Publicação
2023-05-06
Modelo básico
SD 1.5
Introdução à Versão

my first test of mxing model and lora models, it is not bad. The difference between ckpt and safetensors is very small. I may not update it because I'm not majoy in model training. I upload this model just for communication.

Escopo da licença
Fonte: civitai

1. Os direitos dos modelos repostados pertencem aos criadores originais.

2. Criadores originais que desejam reivindicar seus modelos devem contatar a equipe do SeaArt AI pelos canais oficiais. Clique para reivindicar

Escopo da licença de criação
Transmissão ao vivo
Fusão
Permitir download
Licença comercial
Imagens geradas podem ser vendidas ou usadas para fins comerciais
Permitir revenda de modelos ou sua venda após a combinação
QR Code
Baixar App SeaArt
Continue sua jornada de criação com IA no celular