cyno (genshin impact), 1boy, dark-skinned male, male focus, dark skin, animal hat, jackal ears, solo, long hair, sitting, sand, hair over one eye, holding, black headwear, hat, black shorts, shorts
Description
FAQ
Comments (9)
Dude, you are going on a crazy tear with these amazing LoRAs! May I ask how you're processing them? Like are you using a Google Colab or what kind of GPU setup?
hahaha，I just use my own 3070TI video card for training
The selection of material pictures is the most important。
The Lora I trained is used to adjust the face in a real person model.
For me, every time I train an animation lora, I get a new real face model
@kozue I sent a message on discord if that's ok!
Can't generate with this model and keep getting this error. Tried with --no-half and --no-half-vae, but nothing.
modules.devices.NansException: A tensor with all NaNs was produced in Unet. This could be either because there's not enough precision to represent the picture, or because your video card does not support half type. Try setting the "Upcast cross attention layer to float32" option in Settings > Stable Diffusion or using the --no-half commandline argument to fix this. Use --disable-nan-check commandline argument to disable this check.
I don't understand what happened... this is just a regular Lora file, maybe you installed some new plugin?
@kozue according to this thread on GH, it might be related to civit. Are you able to upload this model into another hosting?
https://github.com/AUTOMATIC1111/stable-diffusion-webui/issues/6923
Tried with the latest webui (сommit hash: 22bcc7be428c94e9408f589966c2040187245d81) and only with default plugins - the same error.
@BoberMod Re-download and use, it is normal here……
我来返图啦