Detalhes
Recomendado
V3
V2 - Anime
V2 - Photo
V1
Girls in Glass Jars

Girls in Glass Jars

1.3K
1.0K
5.3K
#Anime
#conceito
#vidro
#tiny
#Onde, cara?
#Realismo
#captured
#tamanho
#Shrink
#bottle
#jar
#Anime
#vidro
#tiny
#Onde, cara?
#Realismo
#captured
#tamanho
#Shrink
#bottle
#jar

Generates images of shrunken girls inside of glass jars! Please share your results, I'd love to see what other people can achieve with this model! Reviews are also always massively appreciated!

Sample Prompt:

<girlsInGlassJars_v3:0.5> (in glass jar:1.45), full body, front view, masterpiece, best quality, 1girl, barefoot, green eyes, long hair, black hair, kneeling, shrunken, detailed, realistic, cork, defined

Common Queries:

"The subject is appearing in front of the jar instead of inside of it."

This unfortunately still happens every so often. Usually regenerating will give you a better result, but if it's happening consistently, you need to add more weight to either the "in glass jar" tag, or the LoRA. I find that (in glass jar:0.45) works well in most cases. Also, adding additional tags such as "trapped", "cramped", "confined", and "contained" can help.

"There are lots of mini-jars appearing inside of the main jar."

This typically means you have too much weight on the glass jar tag. Try removing some. Adding "multiple jars" to the negative prompt can also help prevent jarception.

"Only the woman's head/legs/torso is generating."

This can be somewhat common with certain models, particularly photorealistic ones. Adding and weighting "full body" usually fixes it. You can also generate a good result with an anime model and then use it with ControlNet on your desired model.

"What models work best?"

I get really good results with Corneo's 7th Heaven Mix on clip skip 2, but you should be fine with any anime model.

As for more realistic models, it can be a bit more picky. My recommendations are RealDosMix, Rev Animated, and Deliberate, but it does with with others too.

Ver tradução

Classificação e Avaliação

4.9 /5
0 classificação(ções)

Sem classificações ou avaliações suficientes

avatar
Anúncio
2023-03-20
Publicar Modelo
2023-03-25
Atualizar info do modelo
Detalhes do modelo
Tipo
LORA
Tempo de Publicação
2023-03-20
Modelo básico
Other
Palavra de ativação
in glass jar
shrunken
Copiar
Introdução à Versão

Works more consistently and allows for more prompt variation

There's an occasional issue where feet get turned into glass. Adding "multiple jars" to your negative prompt can help resolve it. I'll be working on a full solution for the next update.

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