@Dr. Furkan Gözükara can you help with one tip
@Dr. Furkan Gözükara can you help with one tip
Pi have 16 images on dataset to train in kohya
Plocally
Pi need keep 40 on repeats?
Pwith 6 epoch its max_train_steps (640 / 1 / 1 * 6 * 2) = 7680
PHow many epochs would you recommend? i using 16GB textencoder config in my 8gb gpu
LI did a new install to setup tensorrt of automatic1111 using your installer and as I recall i've gotten this error before. It looks like I missed something with cudnn maybe. Any ideas?
LI gave up on the tensor I had the same issue before its chasing one problem to another. It's the most broken thing I've encountered. At least with my setup.
Sdid u figure it out?
Popen your path C:\Users\datin\Desktop\New Folder\stable-diffusion-webui\venv\bin
Ptype activate
Pwhen it shows (venv) before path you type pip uninstall nvidia-cudnn-cu11
Pconfirm with y to uninstall
Pafter uninstalled try this pip install nvidia-cudnn-cu11==8.9.4.25 --no-cache-dir
SHi guys, I'm creating a character of a hero for series and want to train sdxl for him as I will use him in a lot of different scenes long term. I created a desired face in MJ (couple of images) and I'd like to pair the MJ face with a body of a real existing man with top physique. (the hero will be mainly without a shirt) Fortunately I have a lot of high res photos (hundreds) of the desired body but the problem is how to pair the MJ face with the body when training SDXL so I'll get consistent full body pictures with the same muscle structure, hair etc. every time in a realistic style. How would you go about it?
P@The MisoGenius i think that in windows you need extract cudnn dll to path in message ...\cudnn\bin you download and extact this files https://developer.nvidia.com/downloads/compute/cudnn/secure/8.9.4/local_installers/11.x/cudnn-windows-x86_64-8.9.4.25_cuda11-archive.zip/
NHello Furkan,
What is the best python version today for Stable Diffusion AUTOMATIC 1111? Can you pls refer me to your most updated relevant tutorial?
Do we use the same python version for Stable Diffusion XL? If not, what is your recommendation? Can you pls refer me to your most updated relevant tutorial?
Thank you Furkan for your kind assistance.
What is the best python version today for Stable Diffusion AUTOMATIC 1111? Can you pls refer me to your most updated relevant tutorial?
Do we use the same python version for Stable Diffusion XL? If not, what is your recommendation? Can you pls refer me to your most updated relevant tutorial?
Thank you Furkan for your kind assistance.
HI keep getting this error whenever I restart runpod & start Stable Diffusion:
ERROR: Cannot uninstall gradio 3.41.2, RECORD file not found. You might be able to recover from this via: 'pip install --force-reinstall --no-deps gradio==3.41.2'.
any ideas why?
ERROR: Cannot uninstall gradio 3.41.2, RECORD file not found. You might be able to recover from this via: 'pip install --force-reinstall --no-deps gradio==3.41.2'.
any ideas why?
NThank you all for your help! Especially @Dr. Furkan Gözükara
Now we're gonna do top content
Now we're gonna do top content
NHello Furkan, I keep getting broken images with artifacts. Pls refer me to a tutorial where you explain how to solve such mismatch of VAE or resolution. Thank you.
NI reinstalled everything from scratch, python, git and stable diffusion 1111, but still getting the same corrupted images.
Nhere is an example
LI think I have it working but what’s the hack to get 768x1344 and 1344x768
Fmy new strategy is
F200 repeat
F1 epoch
Fgetting checkpoints based on n steps
Fyes you will get them
Fthose are not problem
Fprobably several different trainings and find best workingone
F3.10.11 is best i think
Fmost widely supported
Ftry delete venv and start again
FYouTubeSECourses
In this tutorial, you will learn how to install Automatic1111 Web UI for SDXL. How to use LoRAs with Automatic1111 SD Web UI. How to install Kohya SS GUI scripts to do Stable Diffusion training. How to train LoRAs on SDXL model with least amount of VRAM using settings. All of the details, tips and tricks of Kohya trainings. How to do x/y/z plot ...
Fwatch begining 20 min or so
Lyeah but how do you configure a tensor engine for 768x1344 and 1344x768?
LI may have found it actually
Fyou can so easy
Fi have sohwn in video
LI may have it
LI can generate standard sizes but when I change the sliders based in your video to non-standard I get this
Fyes you need multipler of 64
Fthat is limitation
Falso you need to increase max allowed size
FGitHub
I had to modify this code to generate TensorRT that will support highres fix for SDXL self.max_image_shape = 768 if version in ("1.4", "1.5") else 2048 It is by default 1024 So ...
LI tried both but my regular sizes engines are getting errors as well so it's something else. My venv is Python 3.10.6 do I need to update?
PI'm horrified seeing the quality of A1111's python code and its extensions, even more so because I come from the .NET world haha
Fno i just gave you answer
PThis seems like a lawless land
LI don't care how broken it is if it does what I need. I give up on TensorRT I'll come back when the option is less broken