dont use flux1CompactCLIPAnd_Flux1DevFp8
dont use flux1CompactCLIPAnd_Flux1DevFp8
Fuse fp16 base model
Fyes. i also made a video for runpod
Fstill editing it
LThe base model will definitely work on my Nvidea RTX 3060 12gb Ram 32gb video card? Because I can't on the base model to do generation, not enough memory for generation
Fduring training doesnt matter
Fit will be auto casted
Sdo you know when will it be available ?
Fi am editing hopefully tomorrow
SSUPER !
D
D
GLink to site, coupon and the multi GPU video when you drop it. I will do this immedietely
GWhat do you estimate time/cost would be for 1000 images?
MGitHub
A general fine-tuning kit geared toward diffusion models. - bghira/SimpleTuner
LYou tricked me. The 23gb FLUX dev FP16 with preset Rank_5_11498MB_Slow does not fit my 12gb configuration. As before this model is not suitable for training and gening
ADr When I click windows install step 1 it just opens and closes
AAll my other python applications work fine except this one for some reason I use the correct python install 64bit
A
EAre you running your OS with the same 12GB card?
EWhich 1 step installer?
AIn the kohya 1 click installer on his patreon
LYes and RAM 32GB
LI used all Config file train_data_dir. 8 GB GPUs : Rank_9_7514MB.json, 10 GB GPUs : Rank_7_9502MB.json, 12 GB GPUs : Rank_5_11498MB_Slow.json, all give memory errors
PI do this all the time on accident, but make sure you are using the LORA tab and not the Dreambooth tab in Koyha. It will fail with out of memory errors.
EHmm, I can see the 12gb it looks like it is failing pretty early. Do you have a game running in the BG or another stable diffusion UI running in the BG.
It looks like it is failing because there is only 5gb free on the card and it needs 6gb
It looks like it is failing because there is only 5gb free on the card and it needs 6gb
LNo, I don't have any other background tasks, I only work with Kohya.
L
LI don't know, I've been testing everything and something's going wrong.
Fsorry for delay. i was planning actually today too but i got some real life bussines. moving from moutain house to city again
Fsure showna ll
Fdepends on how many epochs
Ftell me i can tell you
Fye i am aware but i still like kohya
Fi train in my rtx 3060
Fhow much vram your system using before starting the script?
Falso dont upgrade torch 2.5 it may use more vram and it didnt bring speed when using single layer strategy
Fplease do this and see what error you get
F
Fopen google remote
Fi will pin point your error
Fif you have 1 time loaded config into dreambooth it will get corrupted
LI haven't updated anything. I have RAM running at max level, before moving text encoders to gpu and after that it gives an error
Fprivate message me your config json file please
Falso show your task manager gpu before starting training
Fscreenshot of it
M1024 Speeds on windows finally working faster 
Fawesome speed
Fused torch 2.5?
Mjust pulled the latest dev one