/vt/ - Virtual Youtubers

>>66019812
Putting the loras in their own sections doesn't actually do anything, really (unless you're on latent mode which doesn't even work with EF?). That's why it's nice to have loras that have good tagging so you get trigger words to apply them to the correct places

Lora in same section as character: https://files.catbox.moe/mgomlx.png
Loras swapped, rest of prompt left the same:
https://files.catbox.moe/34du6e.png

The real thing to be careful with in lora placement in the prompt is that even lora which apply to the entire image (eg the EF+HLL lora) should not be put above the ADDCOMM, otherwise you get very weird results unless you turn the strength way down. eg: https://files.catbox.moe/sqo93a.png