I see, and I think if they trained a 0.9 SDXL version they would have to retrain for 1.0
I see, and I think if they trained a 0.9 SDXL version they would have to retrain for 1.0
Sthanks @Dr. Furkan Gözükara and @ashleyk for response, from what age do you think i should start using ohwx girl or boy? 18yrs and below?
AI wouldn't train minors, Civit AI doesn't even accept models that contain minor content.
Fi think below like 15 age
Fi think he wants to train his personal photo but not sure
Sthanks guys, my intention is to also create an image of my son as a superhero
Mgetting this issue when trying to add the control net extension
extension access disabled because of command line flags
M--share is killing me 
Mbut I think I need
--share to run the webUI interface from runpods terminalMI guess I'll try
--ngrokMyeah ngrok seems to work 
AThis is a security feature, otherwise if you publicly expose the Web UI, attackers can install malware on your Web UI instance using the feature to install extensions.
OLoras are crazy with SDXL. I just trained my first lora last night thanks to Furkan Hocam and getting great results. 
OEven without prompting drawing or coloring book, it gets the style very easily and nicely. Thanks again
TIn the Lora training video (https://www.youtube.com/watch?v=TpuDOsuKIBo) I found the tabs at the top are different (mine say Dreambooth, Lora, Textual Inversion instead of Dreambooth, Dreambooth Lora, Dreambooth TI as in the video)-- more importantly, my Lora tab doesn't have instance prompt and class prompt in it (pretty much everything else is there). Those two fields only show up under the Dreambooth tab (not the Lora tab)-- I'm guessing my version might be newer and the UI has changed, but I'm not sure how to set the instance and class prompt to the 'ohwx man' thats used in the video 
YouTubeSECourses
#Kohya SS web GUI DreamBooth #LoRA training full tutorial. You don't need technical knowledge to follow this tutorial. In this tutorial I have explained how to generate professional photo studio quality portrait / self images for free with Stable Diffusion training.
Our Discord server
https://bit.ly/SECoursesDiscord
If I have been of assis...
Our Discord server
https://bit.ly/SECoursesDiscord
If I have been of assis...
T@Dr. Furkan Gözükara sorry forgot to tag
MI finally got something working but my results are still pretty rough compared to clipdrop's uncrop
Many thoughts one what I can try adjusting in my params?
MI'm also curious if anyone suggests a different sampling method other than
Euler a
Fhello welcome
Fhere more up to date interface tutorial
FYouTubeSECourses
How to install #Kohya SS GUI trainer and do #LoRA training with Stable Diffusion XL (#SDXL) this is the video you are looking for. I have shown how to install Kohya from scratch. The best parameters to do LoRA training with SDXL. How to use Kohya SDXL LoRAs with ComfyUI. How to do checkpoint comparison with SDXL LoRAs and many more cool stuff.
...
...
Fthis one is for sdxl but you will see tabs in this one
Tthanks ill try that
Falso sdxl much better than sd 1.5
Fbut requires more vram
Fwhat is it that your doing ? i dont know clicpdrops uncrop :d
Ti think the base models i want to use are on sd1.5 unfortunately, but i'll try use the video to figure out interface
Flook here : https://github.com/FurkanGozukara/Stable-Diffusion/blob/main/Useful-Commands.md#download-civitai-model
GitHub
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...
SThanks for the help so far, this is progress for me. Prompt for first 4 images is "ohwx woman, photo portrait, in tomer hanuka style" and for the 2nd 4 images is "a photo of ohwx woman" ,used Automatic1111 locally to generate. The last image shows training settings for 10 images. Is this good or can be improved?
Flooking pretty cool. you are using shivam colab?=
SShivamShriraoDreamBooth
Hhi guys, i am following niko's tutorial for running dreambooth but i am getting trainer is not defined error. any idea what that is?
Freeally hard to improve there
Fdo you have bigger google drive spacE?
Fplease follow my tutorials :d
FYouTube
Welcome to Software Engineering Courses (SECourses) – the ultimate destination for skillfully curated insights into state-of-the-art technologies and programming paradigms. We demystify the realms of Artificial Intelligence, Stable Diffusion, DreamBooth, LoRA, ControlNet, Textual Inversion, Software Engineering, Programming, C#, .NET, ASP .NET, ...
S@Dr. Furkan Gözükara yes i have space
Fok then you can do more frequent checkpoint
Fand compare checkpoints
SBy "more frequent" you mean increasing training steps?
TThis is my first time actually pressing the training button in kohya_ss gui (after following tutorial), but i get "RuntimeError: Error(s) in loading state_dict for AutoencoderKL:
Unexpected key(s) in state_dict: "encoder.mid_block.attentions.0.to_to_k.bias",
"encoder.mid_block.attentions.0.to_to_k.weight"... " any ideas?
Unexpected key(s) in state_dict: "encoder.mid_block.attentions.0.to_to_k.bias",
"encoder.mid_block.attentions.0.to_to_k.weight"... " any ideas?
Tah maybe it is because of my base model-- switching to the same one from the tutorial works (realistic vision). The one i was using before fails after loading u-net (or between loading u-net and loading vae)
T
Fthere is save checkpoints steps count
Fyou lower it
Fyou have folder setup error
Fplease watch tutorial again
Tthe folder works fine for using RealisticVision as the source model though, as soon as I change to a different source model, it gives me this error- do some models need something special? (i kept all settings the same between the two runs)