im running 22.2.2 i got no issues but wondering if its better to update
im running 22.2.2 i got no issues but wondering if its better to update
BNot sure what the issue was. Installed a fresh A111 and the problem went away
C@Dr. Furkan Gözükara would google colab produce a better model than dreambooth automatic1111 with my 4090?
F1 hour more to go
Fif working i think keep as it is
Fit depends on the current status of the extension
Fi think we may say yes
Fbecause dreambooth extension of automatic1111 still getting developed and bugs getting fixed
Fbut would it worth hassle? probably no
Fi suggest you to train without xformers, use ema, use low learning rate
Fcompare checkpoints
Ufriends, may I ask if anyone update dreambooth extension in automatic111 webui but the Lora doesn't show up...?
UIt's still like this, but the dreambooth extension is the latest
HNice. I used runpod all day today. It’s very fast
Ufixed, I didn't open in seting LOL sry
MAnyone got a Web UI of ControlNet with Depth they can share with me for testing? I can't do a local install for a few weeks and need to make prpgress on a project for this children's app im working on for free. I need to test out some prompts.
No/ hey all, was follwing the controlnet video and Mikubill's link is broken for the ControlNet extension, dont see anything pinned here, any idea where i should go?
MUsing depth model on ControlNet. Can anyone advise on prompts required to maintain the central character. I did a picture of myself and prompt was "Under the sea background" and it took over my face as well.
Pi would also like to know this.
Pi just watched a yt about connecting the words better and he suggested underscores. doesn't work always, but helps.
MSame thing just happened to me. Last second adjustment in colour. Completely washed the image with green
MI'm struggling to preserve the foreground character and only edit the background without inpainting. I'm trying to automate the process. Inpainting is a manual step
B@Pepper @Manicipal I installed a fresh copy of A111, it notified me of conflicting Python versions during install and recommended uninstalling the other ones. I did that and it's working fine in the new install. I'm not sure if the old folder is working now, I'd have to check in a bit
Ptysm!
MAh amazing. Great
MThanks
MI guess I'm looking to preserve the foreground as much as possible
Rcan someone help me
RI have followed everything step by step in the Dreambooth tutorial, but when it starts generating the samples, all images are black. I have read in the comments that I have to downgrade cuda, so I did it by following the other tutorial on this channel, but I still get black images. any idea? I have a RTX 3090
JWhat cuda version do you have?
RI did the downgrade so now I have this
Jtry this in cmd
Jpython
import torch
>>> torch.cuda.is_available()
Jcheck if it returns true
Rin the cmd windows of Automatic 1111?
JNo, windows prompt
RI mean, there is a cmd window opened, from 1111
Rdo I open a new one?
JYou can open a new one 
JYes
R
JSorry, remove the “>>>”
JJust Python, Then import torch, then torch.cudaetc
R
JSo, this means that cuda is not instalei
Rbut I installed it
JInstalled *
RI followed the other tutorial
JTry this:
J(are you using windows or linux?)
