continue step 2
continue step 2
Fit fixes it
Ldo you know if joy captions happens to apply some sort of automatic 'shuffling' with it's descriptions? Or would applying shuffling be a seperate step?
Sthank you!
Fkohya has it
L@Dr. Furkan Gözükara Did you use bucketing enabled with your style training? I remember you talking about it but I don't remember where you talked about it.
L(I think you said it hurt the quality of the result in the case of getting a person's likeness but I don't think there was a consensus on styles)
Fi didnt all
Fwas 1024x1024
Fgrid is being generated
LGreat I'll keep it off for mine and stick to 1024, thank you!
IHow to use realvisxl in sdxl training
IIf anyone have link please
IRealvisxlv4.0
IThere is only sdxl base model in a1111
Fjust download it
FYouTubeSECourses
Notebook updated to V5 with huge improvements. If you are having trouble to download models onto your cloud platforms such as RunPod, Google Colab, Kaggle, Massed Compute or anywhere, this tutorial is made for you. Moreover, you can upload your files / models into your private Hugging Face repository and save them permanently privately and downl...
II am using kaggle
II haved used sdxl base before
II wnat to try with realvisxl
Fi know that is why i linked you
Fwatch that
Fi am afk
IOk sir thank you
WGood morning! I am doing the KOHYA flux training according to your video. I have an RTX 4090 however I'm only getting 4.39it/s even with the torch 2.5.... Any ideas on what I am doing that might cause it to be so slow?
Wrunning windows 11
FWhich config?
WRank_2_27700MB_4_85_Second_IT
Wwait
WRank_3_18950MB_9_05_Second_IT
WGPU Memory is at 21GB
W@Dr. Furkan Gözükara
![Dazzastrous [4090]](https://cdn.discordapp.com/avatars/871389071686135838/c4a5f4d81d9eeb18d856295d692e2c07.webp?size=40)
Derm the first prompt wheres the airplane 
FGreat speed for rank 3
FUse rank 4 for faster
![Dazzastrous [4090]](https://cdn.discordapp.com/avatars/871389071686135838/c4a5f4d81d9eeb18d856295d692e2c07.webp?size=16)
FPrompt was wrong fixed it :))
Wgotcha. Thanks! Your video was very detailed and informative.
FWill update grids
W@Dr. Furkan Gözükara Using Rank_4_16960MB_Fast.json and still getting
W12/2800 [00:41<2:42:26, 3.50s/it,
FWait 150 steps and let me know
DFBut it is already like 2x of a6000
W@Dr. Furkan Gözükara After 150 steps these are the stats: 150/2800 [08:46<2:35:08, 3.51s/it, avr_loss=0.393]
Scan I use flux dev with 12GB vram?
Fok it is a good speed but slower than linux i guess
FA6000 gets like 4.5 second it on linux
Fyep
GFinal question I just loaded up a 4GPU L40S and going to train, I did 4GPU config, but learning rate is 0.00005? Change it to 0.0001? Your config comes preloaded at 0.00005
Fchange 0.0001
