Krea2 high_heels
Lora type:Krea2
triger：zithighhls, high heels,
prompt：zithighhls, high heels, solo, pantyhose, sitting, legs, lower body, black footwear, skirt, water, brown pantyhose
genarated checkpoint: krea2TurboOfficialComfy_krea2TurboInt8.safetensors
training checkpoint: Krea 2 Raw
recommend strength: 0.9
sampler name: ddim
scheduler: sgm_uniform
ZIT high_heels
Lora type: ZIT
trigger：zithighhls, high heels, feet, toes, sandals, legs, solo, toenails, foot focus, crossed legs, shadow, toenail polish, strappy heels, indoors, close-up, sitting
genarated checkpoint: z_image_turbo_bf16.safetensors
training checkpoint: z_image_turbo_bf16.safetensors
recommend strength: 1
denoise:1
Description
FAQ
Comments (2)
用起来好困难，我从光辉模型过渡到z-image。我不知道该如何配合，光辉模型的两个lora加两套提示词放上去就解决了。z-image按这个方法，感觉自然语言都被污染了，我不知道是不是得用某个节点把这俩的lora自然语言混合在一起才能解决
你应该需要分区提示词功能，不同的lora放在不同的工作区和单独提示词区域，这样可以避免所有的关键词挤在prompt，