Software Engineering Courses (SECourses)•16mo ago

@Dr. Furkan Gözükara I noticed all your generations have very clear backgrounds, how is that possibl

@Dr. Furkan Gözükara I noticed all your generations have very clear backgrounds, how is that possible? Mine always come out with a shallow depth of field which is soft and blurry background

SSebastianblood @Dr. Furkan Gözükara I noticed all your generations have very clear backgrounds,...

Furkan Gözükara SECourses•9/15/24, 1:06 AM

i just replied back on youtube :d

Furkan Gözükara SECourses•9/15/24, 1:06 AM

the LoRA training making that impact. As you do train more epochs, that annoying "shallow depth of field which is soft and often blurry" effect is gone. But more training also causes overfitting so you have to have a very good config + decide which checkpoint to use for balance both

FFurkan Gözükara SECourses stable diffusion mods are out of control

Maxivi•9/15/24, 1:25 AM

didnt one of the mods say its fine earlier ??

MMaxivi didnt one of the mods say its fine earlier ??

Furkan Gözükara SECourses•9/15/24, 1:29 AM

there is 1 mod i thin he is responsible

Furkan Gözükara SECourses•9/15/24, 1:29 AM

so annoying

Maxivi•9/15/24, 1:29 AM

i think thats one of your biggests posts too

Maxivi•9/15/24, 1:30 AM

lot of comments seemed happy

FFurkan Gözükara SECourses the LoRA training making that impact. As you do train more epochs, that annoying...

SebastianbloodOP•9/15/24, 1:34 AM

Thank you

MMaxivi i think thats one of your biggests posts too

Furkan Gözükara SECourses•9/15/24, 1:36 AM

yep huge upvote

Furkan Gözükara SECourses•9/15/24, 1:36 AM

it was one of the most viral posts

FFurkan Gözükara SECourses Click to see attachment

SebastianbloodOP•9/15/24, 1:39 AM

Oh no, damn! This was the flux post with those many awesome images ?

SSebastianblood Oh no, damn! This was the flux post with those many awesome images ?

Furkan Gözükara SECourses•9/15/24, 1:40 AM

yep

Furkan Gözükara SECourses•9/15/24, 1:43 AM

time to do some fine tuning

FFurkan Gözükara SECourses time to do some fine tuning

Maxivi•9/15/24, 2:04 AM

just got word that sd3 8b will be out soon, no distilation and easier to train and finetune

MMaxivi just got word that sd3 8b will be out soon, no distilation and easier to train a...

Furkan Gözükara SECourses•9/15/24, 2:07 AM

licence?

Maxivi•9/15/24, 2:07 AM

i think same as the the updated sd3 medium one with more clarity

Maxivi•9/15/24, 2:07 AM

lykon mentioned

Maxivi•9/15/24, 2:07 AM

looks like it can be better than flux

Furkan Gözükara SECourses•9/15/24, 2:08 AM

nice

Maxivi•9/15/24, 2:08 AM

https://x.com/Lykon4072/status/1834917552209744318

Lykon (@Lykon4072) on X

@linaqruf_ @oron1208 Wait for 8b. It's basically Flux without distillation and heavy hands dpo. This should make it easy to finetune (and dpo).
We're also trying a new scaling down mechanism for mmdit, the new 2b is gonna work much better.

Twitter

•

9/14/24, 11:30 AM

Furkan Gözükara SECourses•9/15/24, 2:20 AM

full fine tuning

MMaxivi https://x.com/Lykon4072/status/1834917552209744318

Furkan Gözükara SECourses•9/15/24, 2:20 AM

nice

Furkan Gözükara SECourses•9/15/24, 2:20 AM

it is also fp16 atm

Furkan Gözükara SECourses•9/15/24, 2:20 AM

not even fp8 it will be lower

FFurkan Gözükara SECourses full fine tuning

Maxivi•9/15/24, 2:22 AM

you can try training on a a100 to see if results are good first

MMaxivi you can try training on a a100 to see if results are good first

Furkan Gözükara SECourses•9/15/24, 2:28 AM

not needed

Furkan Gözükara SECourses•9/15/24, 2:29 AM

it works perfect on A6000

Furkan Gözükara SECourses•9/15/24, 2:29 AM

fused back pass reduced VRAM usage by far

Furkan Gözükara SECourses•9/15/24, 2:29 AM

also there are block swapping

Furkan Gözükara SECourses•9/15/24, 2:29 AM

so it will work on 24gb too

Furkan Gözükara SECourses•9/15/24, 2:29 AM

but fp8 didnt make any diff

Maxivi•9/15/24, 2:29 AM

but very slow though right?

Furkan Gözükara SECourses•9/15/24, 2:29 AM

for some reason

MMaxivi but very slow though right?

Furkan Gözükara SECourses•9/15/24, 2:29 AM

same speed as lora

Maxivi•9/15/24, 2:29 AM

with block swapping?

MMaxivi with block swapping?

Furkan Gözükara SECourses•9/15/24, 2:29 AM

no without

Furkan Gözükara SECourses•9/15/24, 2:29 AM

with block it will be slower i expect

Furkan Gözükara SECourses•9/15/24, 2:30 AM

Maxivi•9/15/24, 2:30 AM

but it wont fit on 24gb without block swapping i think

MMaxivi but it wont fit on 24gb without block swapping i think

Furkan Gözükara SECourses•9/15/24, 2:33 AM

with fp8 it should

Furkan Gözükara SECourses•9/15/24, 2:33 AM

but it didnt work

Maxivi•9/15/24, 2:33 AM

yeah the fp8 models arent making any difference not sure why

MMaxivi yeah the fp8 models arent making any difference not sure why

Furkan Gözükara SECourses•9/15/24, 2:36 AM

just reported it as a bug to the kohya

Furkan Gözükara SECourses•9/15/24, 2:36 AM

it should reduce vram to fit into 24 gb

Furkan Gözükara SECourses•9/15/24, 2:37 AM

cpu offloading almost there

FFurkan Gözükara SECourses just reported it as a bug to the kohya

Maxivi•9/15/24, 4:02 AM

what kohya said

If you specify fp8_base for LoRA training, the flux will be cast to fp8 from bf16, so the VRAM usage will be the same even with full (bf16) base model.**

FFurkan Gözükara SECourses fused back pass reduced VRAM usage by far

dxqb•9/15/24, 7:55 AM

kohya now has fused back pass?

dxqb•9/15/24, 10:21 AM

what did you have for Lora? I'm getting 5.7 s/it using OT, bs1 adafactor fused backpass

FluffyQuack•9/15/24, 11:30 AM

Results from trying to train a character + a puppet mascot character

20240915_1328_Oil_painting_of_a_MSFio_woman_with_orang_flux1-dev_2084106653.png

MMaxivi what kohya said If you specify fp8_base for LoRA training, the flux will be cas...

Furkan Gözükara SECourses•9/15/24, 11:36 AM

he said fp8 doesnt work

@Dr. Furkan Gözükara I noticed all your generations have very clear backgrounds, how is that possibl

Similar Threads