Looking at Flux and training LORAs... Since the text encoder is not being trained, I guess long capt
Looking at Flux and training LORAs... Since the text encoder is not being trained, I guess long captions is not necessary?
PI’ll try thank you.
May I also ask how to get the local url for kohya flux training on runpod?
May I also ask how to get the local url for kohya flux training on runpod?
PI installed everything on runpod and had the link with the config interface for training, but accidentally closed the tab
PJust wondering if there’s a quick way to get the link without running all the install commands again? @Dr. Furkan Gözükara
Fit is coming
Flong not necessary i posted on patreon
Fconnect from 7801 proxy port
Fyou need to run install command 1 time
Fthen you can use fuser to kill running instance and restart
PThank you. So do you think my pod is broken?
P
PBecause when I input the path, no settings are changed
PAm I supposed to also click one of these icons? I’ve tried clicking them all
Ftry button first and last
Fwhat happens?
F
P
PSorry for the phone recording
Fok what happened?
Fyou put accurate path right?
PNothing, no setting have changed
Fpath looks correct
PYes, definitely the correct path
Fverify that you didnt overwrite config file
Freupload it
PThat worked, thank you! 
Fgreat
KWill that even fit a consumer card?
Fyep
Flora is as low as 7.5 gb
Fit will be low too
PI’ve had this error both with inputting the config manually, and also the config json upload
Fmake sure flux buttons are checked
Falso flux lora selected
Flook top to bottom
Flora type will be flux
PLike this?
PAnd here
P
Nthose both match what I have Poppy
P@Dr. Furkan Gözükara I restarted everything and it’s now working - must have been a glitch.
I’ll delete my massive error message above as it’s taking up a lot of room
Thank you for your help as always!
I’ll delete my massive error message above as it’s taking up a lot of room
Thank you for your help as always!
Him trying to installl kohya on wsl and i did succesfully install however i cannot run a flux train ... dr furkan can you help please
Hidk what you did on the fix file for windows to make it work
NI'd love to figure that out as well - speed gains look pretty sweet. my understanding of the fix file though is it just updates torch, torchvision and xformers (and modifies that app so it doesn't try to rollback those libraries). you should still be able to run without doing the fix step, it would just be slower and use more vram.
KI was planning on installing JoyCaption, but I'm looking in the Windows_Install and it doesn do a git clone from anywhere. It only creates a venv where the batch-file is launched. Usually there is a git clone first, then a directory change and then the rest
KIt's probably in the HF_model_downloader.py, but this way of installing is way messier than the previous instakllations you've done @Dr. Furkan Gözükara
Ncould this be because just the models need to be downloaded, then Doc wrote the app.py and it's in the zip, so nothing to clone?
FMy training takes 40 hours, what should I change to optimize it? enable full bf16 training.
enable fp8 training.
running training / 学習開始
num train images * repeats / 学習画像の数×繰り返し回数: 61
num reg images / 正則化画像の数: 0
num batches per epoch / 1epochのバッチ数: 61
num epochs / epoch数: 200
batch size per device / バッチサイズ: 1
gradient accumulation steps / 勾配を合計するステップ数 = 1
total optimization steps / 学習ステップ数: 12200
enable fp8 training.
running training / 学習開始
num train images * repeats / 学習画像の数×繰り返し回数: 61
num reg images / 正則化画像の数: 0
num batches per epoch / 1epochのバッチ数: 61
num epochs / epoch数: 200
batch size per device / バッチサイズ: 1
gradient accumulation steps / 勾配を合計するステップ数 = 1
total optimization steps / 学習ステップ数: 12200
