Software Engineering Courses (SECourses)•17mo ago

In my case it was 🙂

In my case it was

TThereiver In my case it was 🙂

Furkan Gözükara SECourses•9/1/24, 9:44 PM

when you increase batch size

Furkan Gözükara SECourses•9/1/24, 9:44 PM

you need to multiply step speed with it

Furkan Gözükara SECourses•9/1/24, 9:44 PM

to get practical speed

ThereiverOP•9/1/24, 9:45 PM

ahh makes sense

FFurkan Gözükara SECourses you are welcome. also this week hopefully i am fully into fine tuning FLUX

Afroman4peace•9/1/24, 9:48 PM

Do you think its possible to finetune a flux model with 12 GB Vram?

ThereiverOP•9/1/24, 9:54 PM

Pretty sure you can yea

AAfroman4peace Do you think its possible to finetune a flux model with 12 GB Vram?

Furkan Gözükara SECourses•9/1/24, 9:56 PM

we will see

Furkan Gözükara SECourses•9/1/24, 9:56 PM

there are some huge optimizations

Furkan Gözükara SECourses•9/1/24, 9:58 PM

i am working on updating flux installer

Furkan Gözükara SECourses•9/1/24, 9:58 PM

didnt notice difference with xformers

Furkan Gözükara SECourses•9/1/24, 9:58 PM

if it uses lesser vram on windows i will switch to that

Furkan Gözükara SECourses•9/1/24, 10:00 PM

i am also going to improve model downloader

ThereiverOP•9/1/24, 10:08 PM

My darkest dungeon lora so far with mostly your settings. Got me back in the game, thanks

ThereiverOP•9/1/24, 10:09 PM

Just training output images, but progress nonetheless

TThereiver My darkest dungeon lora so far with mostly your settings. Got me back in the gam...

Furkan Gözükara SECourses•9/1/24, 10:09 PM

nice

ThereiverOP•9/1/24, 10:10 PM

315 images, increased lr by 100x. Otherwise all settings are rank 4 i think

ThereiverOP•9/1/24, 10:11 PM

dim 128 and alpha 128 are working nicely

TThereiver 315 images, increased lr by 100x. Otherwise all settings are rank 4 i think

Furkan Gözükara SECourses•9/1/24, 10:11 PM

it may cook model but it depends on purpose :d

ThereiverOP•9/1/24, 10:12 PM

yeah probably, just experimenting to see if I can pull it off this way first

ThereiverOP•9/1/24, 10:14 PM

gotta relearn how to check convergence again. think i was using tensorboard

ThereiverOP•9/1/24, 10:15 PM

been over a year since I've done any training

Furkan Gözükara SECourses•9/1/24, 10:34 PM

new flux train models downloader

Furkan Gözükara SECourses•9/1/24, 10:34 PM

speed mind blowing

Furkan Gözükara SECourses•9/1/24, 10:34 PM

it can resume too

Furkan Gözükara SECourses•9/1/24, 11:15 PM

xformers on windows faster than sdpa

Furkan Gözükara SECourses•9/1/24, 11:15 PM

i think for low configs i will use it

Furkan Gözükara SECourses•9/1/24, 11:34 PM

xformers didnt help

Furkan Gözükara SECourses•9/1/24, 11:34 PM

testing torch 2.5 for 12gb

Konstantinas•9/1/24, 11:39 PM

I am trying to finetune Flux, but getting "no metadata" error. Do I need to somehow prepare metadata, is there any guide on that?

druhl•9/1/24, 11:50 PM

Flux is sooo good, that it is soo tough to decide which lora is best when put side to side on a generator

Ddruhl Flux is sooo good, that it is soo tough to decide which lora is best when put si...

Furkan Gözükara SECourses•9/1/24, 11:51 PM

true

KKonstantinas I am trying to finetune Flux, but getting "no metadata" error. Do I need to some...

Furkan Gözükara SECourses•9/1/24, 11:51 PM

it is my this week research hopefully

Furkan Gözükara SECourses•9/1/24, 11:51 PM

fine tuning flux

Konstantinas•9/1/24, 11:52 PM

cool

I just need to use cloth lora and character lora, and combining those degrades quality by a lot. So finetuning is only solution for me now, hence researching it... But good old Lora works perfectly when not chaining

KKonstantinas cool 🙂 I just need to use cloth lora and character lora, and combining those de...

Furkan Gözükara SECourses•9/1/24, 11:53 PM

ye true

Furkan Gözükara SECourses•9/1/24, 11:53 PM

i think fine tuning may fix it as you guessed

Furkan Gözükara SECourses•9/2/24, 12:02 AM

with text encoder entered the area i am not sure about low vram ranks

Furkan Gözükara SECourses•9/2/24, 12:02 AM

i should test them

Furkan Gözükara SECourses•9/2/24, 12:04 AM

i cant decide clip skip 1 or 0 is better

Furkan Gözükara SECourses•9/2/24, 12:22 AM

majestic armor

Furkan Gözükara SECourses•9/2/24, 12:22 AM

way better than older configs

Furkan Gözükara SECourses•9/2/24, 12:42 AM

i think clip skip 0 better - learns more resemblance

Furkan Gözükara SECourses•9/2/24, 12:42 AM

Furkan Gözükara SECourses•9/2/24, 2:10 AM

generating config for every gpu

Furkan Gözükara SECourses•9/2/24, 2:10 AM

man so many options :d

Furkan Gözükara SECourses•9/2/24, 2:10 AM

Furkan Gözükara SECourses•9/2/24, 2:10 AM

i am gonna compare them

Furkan Gözükara SECourses•9/2/24, 2:50 AM

time to make new update of configs

Furkan Gözükara SECourses•9/2/24, 3:08 AM

configs updated

In my case it was 🙂

Similar Threads

Similar Threads

Similar Threads