main
ComfyUI-Krea2-attention-tweak
A tiny patch altering Krea2 Turbo's attention
This patch is based on the LoRA of Beinsezii Krea-2-Turbo-Projector-Scale-LoRA-Diffusers.
I noticed it contained values too similar to the originals (I compared on a merge) to not have for aim to act as a multiplier.
When set at 3.0 I get more balanced results in between more creativity and prompt following.
Languages
Python
100%