If someone can check that I'm not making an error, that will reassure me that the problem comes from
If someone can check that I'm not making an error, that will reassure me that the problem comes from the installation
Ncouple things I noticed within the first couple minutes... probably not an issue, but it's best to use paths without any spaces in them if possible. also, saw you navigate to ...Installer_v20. v21 is latest.
1Is it possible to update the LLM caption prompt in the Joycaption script tool?
Nthe log of my current train is the same as yours up through "Loaded Flux: <All keys matched successfully>. your log then says "prepare split model" and took you minutes, whereas it only took me seconds... I don't know the answer but hopefully someone else reads this and will know lol - other huge red flag is that you're only using 4 gigs of dedicated vram and 4 gigs of shared. that's not right and makes me think even more that you're using your cpu instead of your gpu. you might try Doc's video here: https://youtu.be/adVhm9aI9Gc
YouTubeSECourses
Many people having issues due to inaccurate Accelerate setup. This is how to setup to make it work properly
1-click installers and best configs for Kohya FLUX training :
A Big complete tutorial very soon hopefully : https://www.patreon.com/posts/kohya-flux-lora-110293257
1-click installers and best configs for Kohya FLUX training :
A Big complete tutorial very soon hopefully : https://www.patreon.com/posts/kohya-flux-lora-110293257
Dthx you, i ever do the accelerate config
Dit's crazy
FYouTubeSECourses
Many people having issues due to inaccurate Accelerate setup. This is how to setup to make it work properly
1-click installers and best configs for Kohya FLUX training :
A Big complete tutorial very soon hopefully : https://www.patreon.com/posts/kohya-flux-lora-110293257
1-click installers and best configs for Kohya FLUX training :
A Big complete tutorial very soon hopefully : https://www.patreon.com/posts/kohya-flux-lora-110293257
Di do this befor the vidéo
Dhello 
Fstill stuck there?
Fi told this to kohya and he said set accelerator should fix it
Fsomeone else had same error
Di try on more time
Fi bet it is also related to windows 11 it sucks :/
D:/
Ftry if fails open google remote
Fi will connect and check
PThat Lora looks amazing, are you the creator of this Lora ?
DIt seems that the problem is fixed by having restarted a second acceleration, I just launched and "I think" that it is good. I come back to you
Flook at the hands
F
Falmost accurate :d
Na loooong way from SD 1.5 hands that's for sure - but even from SDXL
Fis this perfect ?
Twere you able to test this?
Fnot yet
Fediting vi,deo
Tany suggestions for lora settings? I am trying it
the guy in the post says they did 3000 steps, 64 lora rank, batch size 4 and learning rate 0.004 - that seems very agressive, no?
the guy in the post says they did 3000 steps, 64 lora rank, batch size 4 and learning rate 0.004 - that seems very agressive, no?
Fi think my settings should work better
Tcan you share your latest settings? I've seen a few different ones
Tor should i ask on patreon
FPatreon
Get more from SECourses: Tutorials, Guides, Resources, Training, FLUX, MidJourney, Voice Clone, TTS, ChatGPT, GPT, LLM, Scripts by Furkan Gözü on Patreon
Fwhat is your gpu
Fvideo editing almost done
Fi had 4 hours sleep today but slept like 2 hours few hours ago
F
Fexporting video
Fgot to sleep now hopefully tomorrow publishing as first thing
S
Fvideo ready
Fwhich thumbnail is best guys for flux lora training tutorial
Deither 1030 or 1037 maybe?
Fty i need top 3
Fwill do comparison 
Jsecond line #6 or #2
Fty
�These are my picks, because the text is more clearly visible as thumbnail and because they look cool too
JSomething I’m seeing is, why my local trained Loras are like 1gb files and the civitai ones are like 20mb?