Hi again, here's the video of the process I recorded: https://youtu.be/Jk0Ks0N90UU?si=6RRvHfeV-gT9GA
Hi again, here's the video of the process I recorded: https://youtu.be/Jk0Ks0N90UU?si=6RRvHfeV-gT9GAl1
FAutomatic1111 Stable Diffusion Web UI
SDXL 1728x1344 pixels - 30+30 steps
torch: 2.1.1+cu121 • xformers: 0.0.23
vs
torch: 2.0.1+cu118 • xformers:…
SDXL 1728x1344 pixels - 30+30 steps
torch: 2.1.1+cu121 • xformers: 0.0.23
vs
torch: 2.0.1+cu118 • xformers:…
Fyes you cant use gradio share
Fchecking right now
Fwhy you do not use my auto installer?
Fgithub readme file is not up to date
FFYouTubeSECourses
With following this video you can install and use famous Automatic1111 SD Web UI DreamBooth extension on RunPod for full DreamBooth training. Installer scripts are here : https://www.patreon.com/posts/runpod-auto-of-84716845
Click to download scripts
https://www.patreon.com/posts/runpod-auto-of-84716845
SECourses Discord To Get Full Support...
Click to download scripts
https://www.patreon.com/posts/runpod-auto-of-84716845
SECourses Discord To Get Full Support...
AI want to show my colleagues how to use dreambooth and don't think that's a good idea to use your auto installer since they are not you patreon subscribers
Fwell the extension developer stopped updating the extensioın
Ftherefore i didnt update github
Fbut let me update it for you
Athat would be amazing!
Awhat extention do you mean, dreambooth extention?
Fupdated
FGitHub
Stable Diffusion, SDXL, LoRA Training, DreamBooth Training, Automatic1111 Web UI, DeepFake, Deep Fakes, TTS, Animation, Text To Video, Tutorials, Guides, Lectures, Courses, ComfyUI, Google Colab, R...
Fyes extension developer
Fnot maintaining anymore
Aoh, but it's still working, just not updating, right?
Fworking specific commit
Fi modified readme file for them
Athanks so much, will try again now!
this is the only line that was changed, right?
this is the only line that was changed, right?
AUnfortunately it didn't work, result is the same. I recorded the process as well:
https://youtu.be/i0h0uRsKPiY?si=z9EkSvhzM8fGW5E5
https://youtu.be/i0h0uRsKPiY?si=z9EkSvhzM8fGW5E5
Fsome important test
FMedium
Full TensorRT Tutorial is here (42 minutes, 32 chapters) : Double Your Stable Diffusion Inference Speed with RTX Acceleration TensorRT: A…
Fyou were right
Fwe had forgatten 1 part
FGitHub
Stable Diffusion, SDXL, LoRA Training, DreamBooth Training, Automatic1111 Web UI, DeepFake, Deep Fakes, TTS, Animation, Text To Video, Tutorials, Guides, Lectures, Courses, ComfyUI, Google Colab, R...
Fi added it
Athanks 
will try again in the morning and let you know
will try again in the morning and let you know
PWhat GPU driver are you using?
PDose this only work on a fresh install or can I just update my A1111 if so what is the best method to update?
୧Soo stable diffusion randomly generated something NSFW (I wasn't even prompting for a new image, I was face swapping using https://github.com/Gourieff/sd-webui-reactor) with about 5-7 CFG/guidance, and I got insta banned from Kaggle, like within 3 seconds
୧(None of the images I used were nsfw either)
Fit works in all
Fmy driver is
FNVIDIA-SMI 546.01 Driver Version: 546.01
Fthey ban nsfw immediately
PI just delete my SD folder and then reinstall using your script for a fresh install?
Fsure you can do that
୧Is it a good idea to put nsfw terms in negative prompt? There's just no guarantee that something won't pop up randomly if you're prompting for a person or generating variations of an image with a person in it
LCan it be helpul to caption regularization images?
Fi dont think so
Fbecause we use man as a class
Fso we train with man
Fin reg images
LI think your right, I dropped captions just for the hell of it. I think this training regardless will get closer than I have on realistic. Used your MAN class images, using SDXL base and using 1024X1024 size. I have done all my training in 768x1344 because 90 percent of my needs are Mobile Portrait but I found training square and prompting portait worked better. I had less stretched heads and such
FYou could do the merge trick, train one model in 1024x1024 and another in 768x1344 and merge them