Hey. Is it now best to use sdpa instead of xformers for both lora and fine tuning? Doing both flux a
Hey. Is it now best to use sdpa instead of xformers for both lora and fine tuning? Doing both flux and sdxl
ïżœ
Fi feel almost same but xformers just a little bit better
ïżœThanks Dr
PI have two questions about DreamBooth training:
Thanks!
- Is it possible to install Super (super resolution upscaler) on MacBook? As far as I understand, Super can be run on macOS, but it is important to note that some features may require CUDA support (which is limited on MacBook). If you have already used it on Mac, please share your experience!
- Is it possible to do a full SDXL Dreamboot or Dreamboot FLUX training via Google Colab? Or, in your experience, is it better to use cloud services such as Massed Compute or RunePod (I heard recommendations from Doc)?
Thanks!
JHello Furkan,
Here is JP. I recently contacted you on YouTube to report a difficulty that I have performing a with Dreambooth Flux training with Kohya and you suggested to contact you through discord, thus here I am. Let me just first say that I really appreciate what you are doing here for all of us noobs interested in AI creation. I am a 60 years old doctoral program manager (with also several PhD students coming from Turkey!) in chemistry and I only discovered recently this interesting area of AI picture creation. How fascinating! But without your great tutorials, I would certainly not manage! I watch them again and again until I finally manage to solve my problems.
Here is JP. I recently contacted you on YouTube to report a difficulty that I have performing a with Dreambooth Flux training with Kohya and you suggested to contact you through discord, thus here I am. Let me just first say that I really appreciate what you are doing here for all of us noobs interested in AI creation. I am a 60 years old doctoral program manager (with also several PhD students coming from Turkey!) in chemistry and I only discovered recently this interesting area of AI picture creation. How fascinating! But without your great tutorials, I would certainly not manage! I watch them again and again until I finally manage to solve my problems.
JComing to the latest problem, I do not manage to create any Flux dreambooth with your Kohya installation on my PC, unfortunately. I was already successful with Kohya Flux Lora trainings, also with OT Finetuning SDXL trainings, all loca trainings on my PC, but now there seem to be a problem (crash) at the moment Kohya wants to save the first safesensors file. I am running the training on a RTX4080 super with 16GB VRAM (could not afford more, unfortunately), choose your 16GB config (which leads to a VRAM usage of about 14,4 GB, shared usage of 18,3/31.5). Speed is approximately 5s/it, which is ok, I think. I tried with different number of steps (200, 100 âŠ.down to much less), had started with 124 pics (1 repeat), am now down to 15 to for testing purpose. Everything goes fine until Kohya tries to save the first intermediate epoch/safetensor file. Then it crashes, terminates the training, as if it could not save the file. I tried to have it saved at many different places, it makes no difference. Could it be that the size is too big? As already mentioned, I had no problem with the Kohya Flux Loras and OT SDXL Finetunings.
JWhen Kohya tries to save the first safetensors file, it does not say anything, but just crashes. My screen goes dark for a couple of minute. Then I can again check what happened. The console says that the training has ended. I can end you the full log, if you want to try to see what happened. It starts with â--- Logging error ---
Traceback (most recent call last):
File "C:\Python310\lib\logging__init.py", line 1104, in emit
self.flush()
File "C:\Python310\lib\logging__init.py", line 1084, in flush
self.stream.flush()
OSError: [Errno 22] Invalid argument
Call stack:
File "C:\Python310\lib\threading.py", line 973, in _bootstrap
self._bootstrap_inner()
etc. â
Traceback (most recent call last):
File "C:\Python310\lib\logging__init.py", line 1104, in emit
self.flush()
File "C:\Python310\lib\logging__init.py", line 1084, in flush
self.stream.flush()
OSError: [Errno 22] Invalid argument
Call stack:
File "C:\Python310\lib\threading.py", line 973, in _bootstrap
self._bootstrap_inner()
etc. â
JFollowing the advices of ChatGPT, I check for critical errors in the âWindows Event Viewer Logsâ and found indeed several ones at the time of the crash. I can send you the evtx file, if it helps. I also looked into the âreliability monitorâ and there you can also see warnings and application failures at the crash time. What can be the reason?!
Thanks for your help in advance!
Greetings from Berlin
JP
Thanks for your help in advance!
Greetings from Berlin
JP
Fyou mean SUPIR i suppose. our app works on windows but i have no idea can be run on mac which i dont think so
Fbut you can use on even a free kaggle
Fyou can do SDXL on kaggle for free
Fon colab look kaggle code
Fand you can run on paid colab
Fhi i would recommend this
Funinstall your python and reinstall as shown in this video
Fand reinstall kohya too
Fare you using our kohya installer,?
FYouTubeSECourses
If you are interested in using AI, generative AI applications, open source applications in your computer, than this is the most fundamental and important tutorial that you need. In this tutorial I show and explain how to properly install appropriate Python versions accurately, how to switch between different Python versions, how to install diffe...
Pand tell me where to find kaggle code? my main task is to make a photo of the generation with the face of a real person, but replace the body with another. Is there an option how to do this? gpt chat tells me that I should do a full training of dreambooth, then somehow place it in 3D, longer modify or replace the body. What do you advise? Also, I will need generations without censorship (18+)
Fforget kaggle for anything NSFW and i dont support NSFW either
Fkaggle will ban your account immediately
PI agree Doc, I am also against it but I was asked to help with this request. If without NSFW, what would be the best way?
Fkaggle works for SDXL and SD 1.5 training
Pbut based on this video https://www.youtube.com/watch?v=FvpWy1x5etM I concluded that Flax Fal Fine tuning has much better detailing. If you compare SDXL and SD 1.5, the XL version will win in detailing by a lot. Honestly, I am very grateful for your work and contribution to the development of neural networks.
YouTubeSECourses
If you want to train FLUX with maximum possible quality, this is the tutorial looking for. In this comprehensive tutorial, you will learn how to install Kohya GUI and use it to fully Fine-Tune / DreamBooth FLUX model. After that how to use SwarmUI to compare generated checkpoints / models and find the very best one to generate most amazing image...
FFLUX is 10 times better than SDXL
GHi everyone and @Furkan GözĂŒkara SECourses
I have a paired dataset with images before and after retouching (color correction, defect removal on face and body, skin color amelioration). I want to train a generative model to automate this process at scale.
Would Pix2Pix (GANs) be the best choice, or are diffusion models, flux , vision transformers, or other supervised methods better for high-fidelity, production-ready retouching?
Looking for insights on the most effective approach. Thanks!
I have a paired dataset with images before and after retouching (color correction, defect removal on face and body, skin color amelioration). I want to train a generative model to automate this process at scale.
Would Pix2Pix (GANs) be the best choice, or are diffusion models, flux , vision transformers, or other supervised methods better for high-fidelity, production-ready retouching?
Looking for insights on the most effective approach. Thanks!
Fsadly i dont know what can be best way for automating atm
UHi @Furkan GözĂŒkara SECourses ,
I need some advice as I have been struggling with Lora training for a product. I tried with and without captions (without was worse).
I need some advice as I have been struggling with Lora training for a product. I tried with and without captions (without was worse).
Problems I'm Facing
- Trigger words don't work consistently. For example, if I use "KOKUMKF2512NVYP2," sometimes I get some kind of device (see trigger_word_image) and sometimes a dress but with different colors and the same pattern.
- When I use prompts where I explain the dress and background, I get a blue dress, but each time the dress color appears in different shades.
- Hands are distorted a lot.
- When I prompt it to generate on different body types, it doesn't work.
No Caption Lora:
Captioned Lora (see screenshot for dataset):
Mixed Lora of cloth and model head (see screenshot for dataset):
UThank you so much for doing amazing research and sharing with us. And Sorry for the little long message I just want to share as much detail as possible. And is full finetuning is a good choice for this specific use case or not?
Here is some more details:
For the dataset:
Dress: 16 images
Model/Person: 9 headshot images generated by comfy consistent character workflow
I also tried using both lora (dress and person) to create images in SwarmUI but results are disappointing. The major importance for me is to get the pattern of dress, color of dress and person wearing wearing that dress. Please let me know If you need any other details. Please guide me to the right direction. Thanks again
Here is some more details:
For the dataset:
Dress: 16 images
Model/Person: 9 headshot images generated by comfy consistent character workflow
I also tried using both lora (dress and person) to create images in SwarmUI but results are disappointing. The major importance for me is to get the pattern of dress, color of dress and person wearing wearing that dress. Please let me know If you need any other details. Please guide me to the right direction. Thanks again
H@Furkan GözĂŒkara SECourses hey, really enjoyed your videos.
would appreciate a deep-dive comparing the following methods:
would appreciate a deep-dive comparing the following methods:
- full dreambooth on flux (extremely memory intensive)
- dreambooth lora on flux
- full dreambooth on flux + then extract the lora via SVD
- plain flux lora
- plan flux full finetune
J@Furkan GözĂŒkara SECourses I am trying out dreambooth training on runpod using your guide. I got it to run but out of memory on a 4090.
Any guidance would be helpful.
------------------------------------
torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 48.00 MiB. GPU 0 has a total capacity of 23.64 GiB of which 20.81 MiB is free. Process 3037847 has 23.62 GiB memory in use. Of the allocated memory 23.08 GiB is allocated by PyTorch, and 86.61 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables)"
Any guidance would be helpful.
------------------------------------
torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 48.00 MiB. GPU 0 has a total capacity of 23.64 GiB of which 20.81 MiB is free. Process 3037847 has 23.62 GiB memory in use. Of the allocated memory 23.08 GiB is allocated by PyTorch, and 86.61 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables)"
Fhey your trigger word is wrong. it is many multiple words combination
Fi recommend watch this video no skip
FYouTubeSECourses
If you want to train FLUX with maximum possible quality, this is the tutorial looking for. In this comprehensive tutorial, you will learn how to install Kohya GUI and use it to fully Fine-Tune / DreamBooth FLUX model. After that how to use SwarmUI to compare generated checkpoints / models and find the very best one to generate most amazing image...
Fi already did and full dreambooth on flux works as low as 6 GB GPUs
FYouTubeSECourses
If you want to train FLUX with maximum possible quality, this is the tutorial looking for. In this comprehensive tutorial, you will learn how to install Kohya GUI and use it to fully Fine-Tune / DreamBooth FLUX model. After that how to use SwarmUI to compare generated checkpoints / models and find the very best one to generate most amazing image...
Fyes which config you used?
Fand are your images all 1024x1024?
JYes. All images are 1024. The config is 24GB_GPU_22900MB_7.3_second_it_Tier_1.
Mentioning this because I think it's strange. I am noticing when I add the preset to kohya preset nothing changes. I have to go through and manually change everything. and when I save the config. It doesn't save. refreshing the browser will erase all data.
I have 200gb storage, 30 gb of local storage like you say in the video.
Mentioning this because I think it's strange. I am noticing when I add the preset to kohya preset nothing changes. I have to go through and manually change everything. and when I save the config. It doesn't save. refreshing the browser will erase all data.
I have 200gb storage, 30 gb of local storage like you say in the video.
Fthat means you didnt load config accurately
Fit has to load everything
Ftell me your config path you tried to load
JLet me check. this is also what I am running it on. pytorch:2.2.0-py3.10-cuda12.1.
JShould I make sure the folders don't have spaces in the names? I keep remembering runpod doesn't like spaces.
Fyes never use space anywhere
Frunpod starts with
F/workspace/ etc
J/workspace/dreambooth/DreamBooth_Tab_Fine_Tuning_Best_FLUX_Configs/24GB_GPU_22900MB_7.3_second_it_Tier_1.json
JIt looks like the config file path is ok.
Fyes you loaded into dreambooth tab