Thank you for all the high ratings and likes. If you like it, please share the work you have done with this model. Also, please give me a like and high rating if you like! I am not good at English, so sorry if my explanation is wrong.
I made a trial of "LECO", a new training method of "Lora". To learn more about LECO, please visit the following URL.
I personally felt that -4 to 4 is a good weight.
It may not work properly depending on the compatibility with checkpoints and other elements. In that case, I'm sorry.