i am researching 1024x1024 training with kohya atm
i am researching 1024x1024 training with kohya atm
Ffor flux
Jno its so much better, sometimes it happens yes, but most pics are so perfect
Sgood to hear, thanks!
Jdamn my holidays are over now and i cant get enough of this Flux fever 
Stime.... never enough of it 
Jive made alot SDXL images with own loras, really good, but now i just inpainted with flux over the face with the new flux lora and its so much better lol
![Dazzastrous [4090]](https://cdn.discordapp.com/avatars/871389071686135838/c4a5f4d81d9eeb18d856295d692e2c07.webp?size=40)
Danyone tried the Flux Capicitor?
Fkohya brought huge optimizations
Feven 16 gb will fine tune i think :d
Fnot lora
Dso, whats a simple workflow for flux? I see numerous on civitai for comfy, but they have DOZENS of nodes, seems super overkill
Fi am researching right now
Fwith multiple gpu
Ji use SwarmUI and Webui Forge
MI'm looking for anyone who wants to have a website. Right now, I'm providing free website design to boost my portfolio.
Stoday Im for the first time , trying to "Flux"ing it... even I do have high config pc , fp16 beats my patience with loading/unloading, however. fp8 rocks it.. flux undoubtly outclassing sdxl.
cant wait to train it.
cant wait to train it.
VSo Pieter Levels has been talking all over twitter about how he's swapped his photoai app over to using Flux finetunes. Is he straight up just using flux dev for commercial reasons and violating the copyright? Should he expect to get sued or are the companies not really caring about enforcing the license for apps on this scale? https://x.com/levelsio/status/1825213637012893706
SPieter talked to the team behind Flux and obtained a commercial license from them.
AHow can they sue him when they're using copyrighted stuff themselves, allegedly*
Dare you using adafactor??
Fyep i am in researdch
Fyep he is doing 
Fi dont know if flux will go after him
SHe is not doing anything wrong
SHe 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
Fwow nice
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