
Hello! I think there are times when I make the mistake of creating a LoRa when a "new character" has just come out, that makes the training data not have enough and optimal information.
This time I wanted to make a new version, with a lot of information in the training data and VERY carefully selected.
I am very satisfied with the result, I have tested it with many models and the results are positive, obviously if you use realistic models certain details may be lost or come out a bit inaccurate.
Enjoy it and I will be waiting for your beautiful images.
Thanks !
? Instructions for Pony :
Weight : 0.8 - 1 (You can use less weight if you do it with clothes other than your usual clothes.)
Trigger : imtdbellezzz
Standard Belle Prompts : 1girl, short hair, blue hair, hair ornament, hair between eyes, ahoge, medium breast
Standard Outfit Prompts : jewelry, earrings, single thighhighs, skirt, shirt, orange belt, jacket, hood, single glove, orange glove, fingerless glove, socks, grey skirt, shoes
NOTE :
-
The prompts stimulate character generation. They are not essential, they are interchangeable and if you want to omit or replace one, it will not affect the results.
-
Only use the prompts that you want to appear in the image, for example, if you make an "upper body portrait" do not use the prompts "shoes, socks, etc"
----------------------------------------------------------------
Any questions, queries or comments you may have, please do so with confidence, I will read them and respond as soon as possible.
Mi idioma natal es el español, asi que tambien doy soporte en esta hermosa lengua.
If you want to help, I would greatly appreciate it, I am at a point in my life where any financial contribution is more than welcome, a simple dollar would mean a lot to me.
https://www.paypal.com/paypalme/ShutupImTired
or join my Patreon
https://www.patreon.com/GCisalive
描述:
训练词语:
名称: Belle__ZZZ__Pony.safetensors
大小 (KB): 223107
类型: Model
Pickle 扫描结果: Success
Pickle 扫描信息: No Pickle imports
病毒扫描结果: Success