Can I use flux1-dev-fp8.safetensors for training or do I need to use the original?
Can I use flux1-dev-fp8.safetensors for training or do I need to use the original?
Fuse original
JIt was the dreambooth tab. I didn't use the original so I can try that out now.
Fye try it
JThat one strange thing is the config file isn't doing anyhting. very strange to me. It didn't save any of the work from before.
Fclick both
F
Fit has to load values
F
F
JIt seemed to have saved this time. What do I put if I only want to save the 160th epoch in the "Save every N epochs" .
Jthis is what the config file has loaded minus the save every n epoch number. I added 160.
Hamazing --- where are you doing full dreambooth? which script in kohya does that.... not seeing it in the video
Hthere doesn't seem to be any script in kohya that lets you do dreambooth with flux
JI went with the 16gb config file and its working. it looks like there was something that was causing it to fail at 24gb.
I don't know if this matters but I am getting this during the training. however this was also happening during the 16gb training too.
2025-03-22 02:58:21.761881: I tensorflow/tsl/cuda/cudart_stub.cc:28] Could not find cuda drivers on your machine, GPU will not be used.
I don't know if this matters but I am getting this during the training. however this was also happening during the 16gb training too.
2025-03-22 02:58:21.761881: I tensorflow/tsl/cuda/cudart_stub.cc:28] Could not find cuda drivers on your machine, GPU will not be used.
Hthis script will let you do full flux finetuning on low vram GPUs
https://github.com/kohya-ss/sd-scripts/blob/sd3/flux_train.py
but there is no script that does dreambooth training? so it seems like the video you posted is not using dreambooth -- please correct me if i am wrong
@Furkan Gözükara SECourses
https://github.com/kohya-ss/sd-scripts/blob/sd3/flux_train.py
but there is no script that does dreambooth training? so it seems like the video you posted is not using dreambooth -- please correct me if i am wrong
@Furkan Gözükara SECourses
GitHub
Contribute to kohya-ss/sd-scripts development by creating an account on GitHub.
Hunless this line implies using dreambooth: https://github.com/kohya-ss/sd-scripts/blob/sd3/flux_train.py#L88
GitHub
Contribute to kohya-ss/sd-scripts development by creating an account on GitHub.
H
WHi @Furkan Gözükara SECourses I just download your config for SDXL training. For training dreambooth for character how many regulation images from your database you think we need to use train to get good dreambooth character model
SHi @Furkan Gözükara SECourses , I've 33 images of me. can you suggest how many repeats I should consider(config has 40 repeats and 1 epoch, that gives 1320 steps).
Swhat model should I use for the best results sdxl base or any custom model from Civitai (like realvis). Is this the best method for generating images of a specific person?
Fmy config doesnt have save every 160 epooch 
Fsave like every 10 15
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...
Fignore it
Fusing maximum is best
Fso lets say 50 train images and we have 5000 reg images so repeating 100 and make like 2-3 epochs
Ffor 33 images
Fdo repeating 150 and train 1 epoch at least but it may not be sufficient
Fyou can also make repeat 100 and train 2 epochs
UThanks for the response. Here my questions:
- Do you think this will fix the color issue (like the exact shade) and will this also solve different body type problem?
- What can I do so I can try on different body type as well?
- What are your tips on bad hands?
- Do you think full dreambooth training is good choice or not for this use case, as we need exact pattern of the dress?
F1 : get better dataset train on flux
F2 : get better dataset train on flux
F3 : get better dataset train on flux
F4 : well dreambooth is most powerful training atm. better than lora. i cant say it is exactly use case or not but if you want highest quality probably
Fundertanding better dataset only happens through experience. you have to do different trainings compare and improve
UThank you @Furkan Gözükara SECourses . By the way, I am a Python developer and if you need any help regarding maintaining your softwares I would love to do it for you. Thanks again
Fthanks a lot
WHi @Furkan Gözükara SECourses I have another question, can I train multi characters in one dreambooth checkpoint together using different active keywords and caption ?
SShould I use any other model instead of sdxl base?
S- Have you tried the Catvton model ? It takes model and clothing images as input and generates model wearing the input clothing
UNo, I haven't tried. Thanks for sharing. I will try it now. How was your experience with it? Does it work well with different lightning conditions?
SI've tried Simple Catvton workflow. It may not produce exact matching clothing sometimes(color/pattern mismatch). I've seen videos using some advanced workflow to get an exact clothing match. I haven't tried it yet.
SPlease let me know about your experience if you implement it for your usecase
USure, I will update you after trying it. Thank you.
Fthey bleed
FMedium
Thoroughly experimented with Fine-Tuning / DreamBooth training of Flux-dev-de-distill, PixelWave v03, Verus Vision and base FLUX Dev model…
Fof course you can use
Fespecially a model that already can generate similar quality / resembling images of what you want
J@Furkan Gözükara SECourses I am using a 48GB to create some finetuning. I used the 48 config and the learning rate is set at: 0.000004. That seems really low. It's taking about 7 hours on 4500 steps.
WI mean I train using SDXL not Flux i already saw some lora contain many characters