A balanced lora algo ?

[ema7569]

I don't know if it's possible or not, easy or not but my idea is if we add multiple lora in a prompt and under the hood the system merge the results between 0%-100% is to ensure the results are always in 0-100. In other case the image get bad results. This idea is to have an algo that can mix multiple lora weights in a prompt to work automatically in compatible way between all.