Purpose
Convenient VRAM efficient lora for enhancing combat, actions and motion of highly dynamic scenes combined with reasoning.
Works with FL2VA, REF2VA and Hybrid Models.
Credits
An experimental consensus crafted lora of multiple aligned concepts of
Combat
Action
Motion
Reasoning
💫 All credits for the original loras go to FourBunny and MisticRain69
If you want to support this, support these two for their awesome work!
Description
FAQ
Comments (12)
will it work with nsfw purpose?
Not tested
NSFW Combat? A thing to see.
"my imperial father ordered me to crush the rival kingdom's top female knight but we touched swords and accidentally started making babies!" ahh anime title
@Inusanfox Was a thing in different times and places in the ancient world. Most men would have hoped they weren't on the losing side. There were also some wrestling traditions where the winner would mount the loser.
@ss9999 Yes! Yeesss!!!. See. These ARE things to see. Pic's or it didn't happen.
this is useful lora thx.
dude you're on a rampage, thanks for another one
Works with singularity hybrid model?
I already have H3_Combat_V2.safetensors lora, appr 150 mb. On which point is your lora better?
that previous H3 motion combat lora didn't made significant improvement to me. it blurred things even more.
so i'll try it out, then i'll combine both loras and decide for myself
It is not a replacement of H3_Combat_V2.safetensors, it it a blend of multiple loras who would use more like 2GB to enhance the mentioned aspects.