He got a commercial license and contacted them
He got a commercial license and contacted them
Fah in that case makes sense
Fyes i will use adafactor
Fminimum vram so everyone can train on their computer
SThe guys at Black Forest Labs are pretty awesome in terms of licensing and monetization
Fthey are doing everything accurate stability ai failed
DIs training on my rtx 3090 but 9 times slower than sdxl, lora rank 8 1024x1024 resolution and adafactor
D9.61s/it is normal?
Flooks like
Fi will look for more optimized way
Didk but it's relativ. people have posted good results after 2000 steps. sdxl is big mess after 2000
Dhowever:
Fnew app coming
Dis too slow, is using a little bit of shared memory
Fi got a lot of settings to test but taking time and constantly new stuff getting added to kohya
DI'm going to try 960x960 resolution
Dno, that wasn't mine. just a screenshot from an early adopter on another discord
Dohh ok thanks
Dclearly he was using non-consumer hardware, but even the best one usually not 3x faster
DI have that speed at 512x512
Fnew app coming
F
Fsomeone on reddit requested this
Fi need to publish this to continue flux training 
Fmaking kaggle runpod and massed compute installer to publish
Dthe quality compares to supir?
Fi didnt have chance yet
Fapp published
FPatreon
Get more from SECourses: Tutorials, Guides, Resources, Training, MidJourney, Voice Clone, TTS, ChatGPT, GPT, LLM, Scripts by Furkan Gözükara on Patreon
Dlooks nice
Dlets give it a try
FReddit
Explore this post and more from the StableDiffusion community
ADid kohya work out Dr or did it still have bugs?
Fi found the bug
Fhopefully gonna start trainings today. but they keep fixing bugs adn adding new features
Fi am also following every messages since there are so many parameters
Dcivitai is already exploding with flux LoRAs
ACan full size version be trained on 4090?
Fi think yes
Fi recently tested a lot of SDXL lora and most of them are trash :d
HLooking back now my SDXL Lora’s were very bad
FFyep mine too
AWhat made them bad Dr?
Flesser research and lora inferiority
Vwhy using sdxl and sd1.5..flux is here now..waiting for someone to make flux as optimized as sd 1.5 and sdxl so that i can have similar generation speeds as sd1.5 and sdxl....
Fgeneration speeds depends on model parameter size
Fbu for training hopefully i am gonna make 24 gb and 16 gb and 12 gb config
Vdoc what about GGUF Image Models ??
