for 2x you need to run training with --no-half
for 2x you need to run training with --no-half
FYouTubeSoftware Engineering Courses - SE Courses
Our Discord : https://discord.gg/HbqgGaZVmr. This video I am showing how to downgrade CUDA version for proper training and I am showing how to do LoRA training with 8GB GPU.
If I have been of assistance to you and you would like to show your support for my work, please consider becoming a patron on
https://www.patreon.com/SECourses
This CUDA ...
If I have been of assistance to you and you would like to show your support for my work, please consider becoming a patron on
https://www.patreon.com/SECoursesThis CUDA ...
Fhi i have the same error as them and i am confused where do i apply this
Fsorry im new to all this
Pthis worked for me. do this for sure
LThanks !
Fhi let me give you file to replace
FFreplace this file
Fthank you
Fthat worked
Uif i install no half to use v2 model do i have to uninstall it to use v1.5?
UYeah I've already uninstalled it
Ohi @Dr. Furkan Gözükara , just finished your selfie --> AI video. Love the results! But I made my prompt "danface" which I think was a mistake. Can I change the instance prompt without retraining the entire model? It's a good model, I just feel I messed up the prompt. Thank you!
Fno you can use both. --no-half just make it run slower for sd 1.5
Fnope you have to retrain
O@Dr. Furkan Gözükara Thank you! Will do 
UI have already used no half and the samples keep appearing brown
Fwhich model you using
Fsd 2.1 512 or 768
U2.1 512
Fare you 100% sure
Fyour training imges are also 512x512 ?
Uyes
UYes
Fyou tried to generate imges
Ffrom that model
Ffrom txt2img tab
UWhen I use the model to create images, brown images do not appear
UOnly with the samples
Fdo sd 1.5 work currently
Fwith same setup
UWhen I use the 1.5 model I don't have problems with the samples, only with the v2
Fi can quickly connect your computer to cehck
UOk!
LI'm just doing a few tests to gain practice. Why is it that when I ask it to generate an image, it creates multiple subjects (people) despite the fact that I ask it "1girl, only" ?
L
Fprobably related to resolution
Fwhat is your resolution
LOK so I don't have to ask for 1080p images ^_^'
LLast question (for tonight ^_^), about safetensors file where i need to put it ? I've download it but i can't understand the correct folder
Fsame folder as ckpt
KHi guys, I can’t run dream booth
KI am getting this error
KRuntimeError: don't know how to restore data location of torch.storage._UntypedStorage (tagged with mps)
Pipeline or config is not set, unable to continue.
Pipeline or config is not set, unable to continue.
K
Dahh
KThis is my version
KWhat is the error?
Dyou will have to update and remove the dreambooth, and then get the new one
乂which model trying to train?