thanks - right now im doing Wan2.2 training on my dataset ll see how it turns out
thanks - right now im doing Wan2.2 training on my dataset ll see how it turns out
CLet me know please. Share also what datasets you used if you can. Just photo's or mixed or video clips?
Sdue to nda i cant share dataset but these are studio photos and few natural shots of a person. Unfortunetly only photos for now
Mhello, i got out of memory error using massed-compute 2x RTX 6000 ADA
Jtraining what?
M40 images of a human for Flux
SNo LoRA vs LoRA Wan2.2
Sshe is little bit damaged tho its 40 epoch [65img's dataset] 32DIM
Jlearning rate?
Fdid you load lora config into dreambooth or dreambooth into lora?
Fconfig name?
MNo, I loaded dreambooth config into Dreambooth. I tried both 48GB_GPU_28200MB_6.3_second_it_Tier_1 and Batch_Size_7_48GB_GPU_46250MB_29.1_second_it_Tier_1. out of memory errors for both
Fwhat is your images resolution?
Fi recommend restart machine and use this config load into dreambooth
Fwill work 100%
F48GB_GPU_28200MB_6.3_second_it_Tier_1.json
S3e-4
M1024x1024. It works in single GPU, but always memory error when I try to use dual GPU
JI will try 0.000234 next as learning rate
Jjust a little bit more than 2e-4
JAt ~ 1000 steps
JI have no idea if amount of images affects the learning rate, if we keep the steps the same
Smaybeeee
Slet me know 
JSure thing but I think I am on to something 
Fdual gpu dreambooth need 96 gb
For lets say at least 80 gb
Fgpus
Mi got memory errors on 2x RTX A6000 (48GB)
Fyes you cant train dual gpu dreambooth
Fbut you can train lora
Mi see in your training log 256_img_BS_7_Fine_Tuning.txt that you trained 8x GPU dreambooth with NVIDIA RTX A6000 VRAM 48577MB Arch
Fyes but that 8x gpu was used for different trainings
Fnot as a combination 
Flora works with multiple gpu but dreambooth fails unless you get big vram gpus
Jso whats better for fine tuned image gen realism right now?
dev or krea?
pros and cons? can someone give me their observations?
dev or krea?
pros and cons? can someone give me their observations?
Fi posted grids and hopefully will make a video today
Fstill working on wan 2.2 presets
Jwhere can I see your grids, patreon?
Jand how about wan 2.2? kling 2.1 master still better?
Fi dont use kling sadly
Fgirds here https://www.patreon.com/posts/112099700
Patreon
Get more from SECourses: FLUX, Tutorials, Guides, Resources, Training, Scripts on Patreon
F
JI appreciate you taking the time to make that
JI personally don't see any differences/patterns
Jprobably going to stick to flux.dev for now
SHey guys, can confirm the tests of our dearest Dr Furkan.
Flux Krea training is amazing but needs more epoch. I trained a makeup style checkpoints with 200+ images, and this is epoch 200. With flux dev the sweetspot is often around 140/150 epochs for my trainings.
I will train 20 epochs more on this one cause I think it may be beneficial for some details. Go ahead Flux Krea is awesome, reach for high epochs !
Flux Krea training is amazing but needs more epoch. I trained a makeup style checkpoints with 200+ images, and this is epoch 200. With flux dev the sweetspot is often around 140/150 epochs for my trainings.
I will train 20 epochs more on this one cause I think it may be beneficial for some details. Go ahead Flux Krea is awesome, reach for high epochs !
Snice, whats your learning rate?
S4e-06
Fmatching with my results