Whether it is necessary to use vae while training
Whether it is necessary to use vae while training
Fyes ofc
Fbut usually models have them embedded
MOk then. I can use that embedded one ...
Fif you can afford to spend few bucks this is the very best tutorial you can find
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...
MSir do u have specific video about dreambooth sdxl training
FYouTubeSECourses
In this tutorial, I am going to show you how to install OneTrainer from scratch on your computer and do a Stable Diffusion SDXL (Full Fine-Tuning 10.3 GB VRAM) and SD 1.5 (Full Fine-Tuning 7GB VRAM) based models training on your computer and also do the same training on a very cheap cloud machine from MassedCompute if you don't have such compute...
MWhether setting resolution as 512 will affect the model
FHi Doc, I saw the beautifull images of the Cookies you trained on Linkedin, any tipsof training a object rather then a person?
FIm trying to train a coke bottle using your finetune flux method
FI got good results for the bottle alone but when I prompt someone drinking , it usually distorts the bottle
Fgets the logo wrong, or the proportions opf the bottle and the hands are worng
Fthis is the dataset Im using
FI trained on prompt" Coca-Cola" and Class Bottle
Fany tips or advice on how to make it better?
Fthanks
FAs my goal is to get a person drinking Im thinking that maybe instead of fine tune I should train two loras, one with the bottle and one with people drinking the bottle and then just mix them
Fyes 1024x1024 best for sdxl and flux
Fdataset
Fit is all about dataset
Fyour dataset is bad :d
Fif you need random person
Fyou dont need to train person
FWhats bad about it?
Fbut how about the size of the bottle in relation to the hand?
JNot hifi images
JAnd as Doc says, "unwanted noise" on the dataset, like people
Fmake your dataset has relative sizes matching to person holding it in your dataset
FSorry Doc I got confused with this so you think I can include people holding the bottle
Fdo you mean relative sizes to other objects right? no to people
Fno people, people is noise
Fdistant shot of the product
Fso the proportoion size of the product
Fis like same as person holding it
Fgot it
FDoc I'm getting this error no Massedcompute
F
FUsing the 48GB_GPU_28200MB_6.4_second_it_Tier_1
FI noticed the FLux option is gone
Ffrom the model selection
Fwhat am I missing?
F
Fcould it be because I rented a 48g GPU and used the 48g config file
FNope same thing maybe I'm using an depricated config file?
Ddoes MassedCompute only offer Linux? I need a Windows cloud host
FTryed a bunch of different config files, even updated
message.txt16.69KB
Fstill nothing
Fcurrently linux
Fi told them windows as well
Fthey said they may add