English is not my native language, so I used machine translation. I apologize if it makes me hard to understand.
(I think it's pretty bad)
Artist akuma_(st.takuma) Style LORA：
5000 training sessions,
142 data sources,(yes, very few)
Version A:
Version A is the version of the original data (original images) with keywords marked.
Version A Trigger keywords:akuma_(st.takuma)
Version B:
Version B is the version without any keywords.
Version B Trigger keywords:Theoretically, no
Recommended number of steps:10~20
Regarding commercial use:
This LORA is completely free.
As long as the original artist doesn't object, you can use it freely, but please don't attribute your personal actions to me.
Description
FAQ
Comments (1)
Almost forgot The keywords in version A are not mandatory keywords,If a character or action is fixed, you can consider removing it.(Because AI may recognize it as a character or action with a similar name)