V1 is overbaked so use low strength, it's trained with double caption, Blip and WD14 because more is more. No triggerStrength between 0.6-0.8 seems to be best
moreToppingsCaptioned-SDXL.safetensors