Detalles
Recomendado
v2.0
v1.0
KNK PonyMerge

KNK PonyMerge

8.8K
66
2
#Anime
#Modelo Base
#general
#lora

Merge model created for testing my Pony Loras in an easier way. Think about it like an anime friendly lora model, like an Autismm or PonyForAnime, but more flexible and colourful.

This merge doesn't brighten based on their base style; it brightens using it with Loras! Anyway, you can still ???????? interesting results.

Version 2:


I've updated and balanced a little the merge:

  • Now has a little more anime alike style by default (just a little)

  • Better hands and details in general

  • Gens pretty similar results like the v1

  • I updated the recommendations, them works for v1 and v2

Results:

It generates shiny, high-contrast images with excellent Lora compatibility:

Images without any negative tags.

score_9, score_8_up,score_7_up,score_6_up, a girl sitting on a park bench, dutch angle, smile, open mouth, <lora:tanabae_kyou_style_pony6_v1-000014:1>

Full size

score_9, score_8_up,score_7_up,score_6_up, a girl sitting on a park bench, dutch angle, smile, open mouth, <lora:maeshima_ryo_style_xl_v1:1>

Full size

score_9, score_8_up,score_7_up,score_6_up, a girl sitting on a park bench, dutch angle, smile, open mouth, <lora:kink_style_xl_v5:1>

Full size

Full vs Bf16

I have used the full version for my latest Lora post, but I think most people don't want to use 13.3 GB with the media at 6.7 GB. Therefore, I resized the Model to make it smaller, but it still produced good results (with only slight differences in the most cases). Once both are loaded, they use the same amount of VRAM, so you can use whichever you prefer.

Vae

Use the default one, sdxl_vae.safetensors, or if you want more brightness, you can go for XL_VAE_C e7. The results are excellent:

BF16 version still no neg prompt


Full size

Recommendations:

My default prompting configs are the next:

score_9, score_8_up,score_7_up,score_6_up,score_5_up,score_4_up

[Neg]
source_cartoon, source_furry, monochrome

Euler_a 20steps CFG 5-8 and use hires fix when is necessary.

Use one of the next resolutions:

    (1024, 1024),
    (1152, 896),
    (896, 1152),
    (1216, 832),
    (832, 1216),
    (1344, 768),
    (768, 1344),
    (1536, 640),
    (640, 1536)

License

This model follows the same Pony V6 License.

Ver traducción

Calificaciones y comentarios

-- /5
0 calificaciones

Aún no se han recibido suficientes calificaciones o comentarios

avatar
Konoko
25
491
Anuncio
2024-05-06
Publicar modelo
2024-07-28
Actualizar información del modelo
Detalles del modelo
Tipo
Checkpoint
Fecha de lanzamiento
2024-07-28
Modelo básico
Pony
Alcance de la licencia
Fuente: civitai

1. Los derechos de los modelos republicados pertenecen a los creadores originales.

2. Si los creadores originales desean reclamar su modelo, contacten al personal de SeaArt AI a través de canales oficiales. Haz clic para reclamar

Alcance de la licencia de creación
Transmisión en vivo
Fusión
Permitir descargas
Licencia comercial
Las imágenes generadas se pueden vender o usar con fines comerciales
Permitir reventa de modelos o su venta tras la integración
QR Code
Descargar la aplicación SeaArt
Continúa tu viaje de creación con IA en móvil