A mix of many attractive Instagram influencers photos.
This lora gives the Instagram/phone filter effect that all phone do nowadays, like smoother skin, makeup, perfect to make amateur-like photos. Especially good for KREA 2 which likes to add too much skin details.
This lora will also tend to make the girls look like Instagram influencer: big lips with lipstick, flawless tan skin, large breasts, perfect hair, makeup, etc
Based on the same datasets as my popular Instagram Baddies Loras.
For WAN: Trained on Wan2.1-T2V-14B. Works both for T2V and for I2V.
For the Z Image Turbo model: If you're using ComfyUI versions 0.3.76 or more, use the pruned lora version If you still have an older version of Comfy, use the not pruned version. More explanations in this article: ComfyUI Updates Changed My Z-Image Turbo LoRA Output (and How I Fixed It)
š If you like this model, please click the Like button and post some pictures, it really helps for visibility! š
Access all my celebrities LoRAs or request a custom Lora
Description
FAQ
Comments (8)
next human update needs at least three testicles/ovaries
Don't know why but the krea2 turbo lora crash here, maybe it isn't for img but video?
Krea 2 is an image model not a video model. Does your workflow work withou the lora ? maybe you dont have enough VRAM ?
@GgrueĀ It does work without the lora, it doesn't seem to be an issue with this lora in specific, any lora for krea2 don't seem to work here, it just stuck on the Ksampler without making any progress pass 12% perpetually, i was wondering if maybe it is a memory issue since it doesn't give me any errors, but with 24GBvram it shouldn't be a problem right?
@LonmineĀ it can be a problem of vram yes, i had the same issue with 3090.
@LonmineĀ it might not be your "gpu" but your dependencies and stuff. If you make sure to read the error log during comfy UI startup, it might tell you there's an update for pytorch or something of that effect. Generally these "memory" problems when you have enough memory is related to comfy's memory management code. If you feel you're up to date but still seeing that it won't continue, trying different comfyui startup flags like "--disable-pinned-memory" or "--disable-smart-memory" or "--disable-dynamic-vram ". You could try one at a time and see which works, and that might get you closer to the "real problem".
@makiaeveliĀ Thanks, i will try it out later, but i found the issue, it seems i cannot load the text encode together with any lora or it will just get stuck, so i put the text encode to be processed on the cpu and now it works fine, weirdly enough even using vram and ram cleanup and even though it shows the memory got cleaned it is still impossible to change text encoders or models, i have to close comfy and open again to do it, comfy memory management is... complicated it seems.
Krea2 LoRA is Making Plastic faces