I mean manual mask mode
I mean manual mask mode
RWhen desactivate the option "use auto-generated mask"
Fyou are right plese remind me tomorrow if i forget
Fsorry for late replyt
RI think the call to the first element of the array Layer is wrong, maybe it try with the second
Fpublished even better version
FPatreon
Get more from SECourses: FLUX, Tutorials, Guides, Resources, Training, Scripts on Patreon
Fperfectly now you can mask 
Ftook me whole day to code
ᴏJust saw message, thanks. using Dr. Fukans.
AI tried switching to prodigy optimizer setting learning rate to 1 and cosine. With 60 images using basic tag captions, 150 epoch, batch 4 the result was very amazing and fast.
AIt actually overtrained
AI have never used it before though so I will need to fine tune the settings and educate myself on the differences between it and adafactor or Adamw
Fnice give it a try
Fso far i see that optimizer makes minimal difference as long as you are using most optimal parameters. my new workflow will be only slightly better
SSWe have it enabled at mitte.ai already before other platforms and thought you might want to try it first before others
Fsure add my account some credits i can ttest
Sadding!
Sdone 
Fty
RWhat GPU did you use?
A4090ti
AI am having some trouble tbh it is like 80% there but I cannot push it to 100% if that makes sense
AI have tried to increase steps and decrease d coef
Adecreasing d coef helped with theh ands and eyes a little but still getting heavy artifacting in many pictures
Awell, not HEAVY, but visible here and there
Aenough that it is irritating
AMaybe not for me after all
RInteresting. Would you say that Furkan's configuration based on Adafactor still gives higher quality results than with prodigy?
ROr what do you mean by 100%?
Asorry for delayed response, still playing around with it. So far, yes absolutely
Abut the speed increase is undeniable so I will continue to experiment
Athere are other parameters I am testing now (dropout, weight decay, further refining d coef) as well as changing batch size/repeats/epochs to see if I can arrive at a good setup
Ahttps://github.com/derrian-distro/LoRA_Easy_Training_Scripts Also found some other program (not that it will be helpful to train but something to play with)
AI believe it is Khoya based
Abut a different front end similar to how swarm is to comfy
Abut I could be wrong I'll take a look and see if I like it better than onetrainer. Khoya UI I didn't like much so this might be a nice solution
I'll take a look and see if I like it better than onetrainer. Khoya UI I didn't like much so this might be a nice solutionRThank you for expanding on your experiments. Are you using an LLM to guide you through tunning parameters such as dropout or weight decay? I wonder if SOTA LLMs can help us arrive at the optimal ones faster than doing it on our own.
RAre you planning to share your findings? I once tried to fine-tune with prodigy on Kohya but I could not make it work.
ASure if course
AI am not using llm just researching the old fashioned way
AAnd trial and error
Agetting close - in fact I think i have it the only problem I have now is the problem I always have which is with unique features getting lost (moles on specific sides of faces, or heterochromia, etc.)
AI am going to go back and try to weight images which highlight those features more heavily in my training dataset
ABut it might be a day or two actually not feeling well today
RNice. I appreciate it!
L4090ti prototype? sorry that catching for it but I was looking ti without success, as I know they decided not to Release it but I can be in mistake
sorry that catching for it but I was looking ti without success, as I know they decided not to Release it but I can be in mistakeAyeah i thought for sure i had ti
Abut you're right