Can we have a regularization weight like AI tookit reg_weight?

[diodiogod]

Or is there already a parameter similar to it on kohya?
I feel like I got much better results using regularization for some concepts with reg_weight at like 0.25 on AiTookit while with kohya regularizations It always seems to drag the learning way too much.