facechain
facechain copied to clipboard
why facechain train with 512 image on XL ?
when train xl lora as you can see in the app.py train image resolution is 512 is it same with sd1.5
I cant understand why facechain 1024 resolution for train XL lora . can you answer this?
because almost training XL LORA case use 1024 resolution image for train
any reason for use 512 ?