
World Morph | Honeycomb
This is my first World Morph LoRA, and first SDXL model!
I think I need to test it more thoroughly, but hopefully it functions properly as a world morph, otherwise you may want to raise the weight to add more honeycomb influence starting from 0.5 and working your way up.
Training images from Copilot Designer (DALL-E 3), and trained on the Civitai On-Site Trainer.
Recommended Parameters
Weight: 0.5-0.75
Keywords: (any input) made of honeycomb
Sample Settings
0.7 Weight, 12 steps, 4 CFG
ZavyChromaXL8. Generated in ComfyUI, with Perturbed Attention Guidance node at default settings, with 12 hires steps, 2x upscale at 0.56 denoise.
( some samples are being generated with on-site generator )
I recommend using whatever you feel comfortable with, this is just what I used to generate the showcase samples.
Keywording Examples:
house made of honeycomb, exotic dress made of honeycomb, lamp made of honeycomb, kitchen made of honeycomb, gaming chair made of honeycomb, honeycomb this, honeycomb that, etc
Also try to describe the scene /background and maybe add extra honeycomb influence to the prompt to explore more of the World Morph capabilities. Simple prompts may work fine, but it could be more interesting to apply to a more complex scene.
As this is my first SDXL model and World Morph, I'm open to criticism and hoping to see images using the model!
描述:
训练词语: (anything) made of honeycomb
名称: SDXL_World_Morph__Honeycomb-000011_akumakushu.safetensors
大小 (KB): 223101
类型: Model
Pickle 扫描结果: Success
Pickle 扫描信息: No Pickle imports
病毒扫描结果: Success