Hello ♥

for whatever reason you want to show me appreciation, you can: ❤️Ko-Fi❤️

This is my result from experimenting with Pony Models.

This model aims to create slightly more realistic images.

Pretty sure that this model still has weaknesses, but with your help I can of course improve it faster

You can create all kind of images with this Checkpoint, but if you enter nothing, for pony unknown words or an unclear prompt, it always tries to generate women.

You can see some examples in the Model Images.

Attention, the model has an urge to take nsfw pictures.

take in mind that i have informed you :D

Recommended Settings:

Pony Model special feature: (check documentary in Original Pony Model)

Sampling Method Euler A

Sampling Steps 20 or more

CFG Scale 5-7

Upscaler: SwinIR _4x (this is my all time favorite, the most others works well too )

(Hires steps: 15; Denoise: 0.28, upscale by 1.5 - 4 (depends on your GPU))

📌 USE sdxl_vae

you have to add the following trigger in

Positive Prompt

score_9, score_8_up, score_7_up,

Negative Prompt

score_6_up, score_5_up, score_4_up

Additional Trigger (optional in positive or negative prompt)

'source_pony', 'source_furry', 'source_cartoon' and 'source_anime'

I can also recommend using the these embeddings

All Images generated in SD Forge with Nvidia RTX 4090

Please test other Settings and share with the communty, so i will learn about the Settings and can optimize.