I completely changed the tags and the dataset balance.
In this version my LoRA should be way more stable and should adapt well with more models than before.
Now i also added the ability to use or not the iconic red jacket. Just add "red jacket" on the prompt.
As extra you can also generate the demonic version of Charlie, isn't perfect yet, but It's a start. As for the jacket you can swap between Charlie "standard" and demonic using
as trigger "charlie morningstar" or "chariedemonic"






