Software Engineering Courses (SECourses)•11mo ago

faster but i dont know if easier

Furkan Gözükara SECoursesOP•1/31/25, 12:51 AM

i didnt use yet

Fabricatedgirls•1/31/25, 12:52 AM

is there something like kling elements in SD that lets you use reference images without having to train a lora

Fabricatedgirls•1/31/25, 12:53 AM

like if i had a spaceship design or a creature or something

Fabricatedgirls•1/31/25, 12:53 AM

and wanted to use it in SDXL or flux or whatever

Dazzastrous [4090]•1/31/25, 12:53 AM

hunyuan

Fabricatedgirls•1/31/25, 12:53 AM

is that not for video?

Fabricatedgirls•1/31/25, 12:54 AM

or are you not replying to me? Sorry just catching up

Dazzastrous [4090]•1/31/25, 12:54 AM

sorry thought u meant video

Dazzastrous [4090]•1/31/25, 12:55 AM

omnigen maybe?

Fabricatedgirls•1/31/25, 2:06 AM

not just for still images for now

watersocks•1/31/25, 3:10 AM

sorry i see this question has been asked and I seem to be doing it correctly but still getting the "model failed to load" error. Is this in the correct folder?

watersocks•1/31/25, 3:10 AM

on massed compute

daroth•1/31/25, 8:04 AM

its a paper launch

daroth•1/31/25, 8:04 AM

https://www.youtube.com/watch?v=wMd2WHKnceI

YouTubeGamers Nexus

Paper Launch

Sponsor: Antec Flux Pro case on Amazon https://geni.us/CaPv
This video talks about NVIDIA's paper launch of the RTX 5090 in addition to its launch of the RTX 5080 that no one wants. We crawl through the internet to try and learn more about just how low the inventory and stock was, as well as share our own experiences of trying (and failing) to b...

umuthasanoglu•1/31/25, 10:00 AM

Hi guys, which captioning model is the best to use with SUPIR upscaling? Also what's the best configuration settings for SUPIR?

Wwatersocks sorry i see this question has been asked and I seem to be doing it correctly but...

Furkan Gözükara SECoursesOP•1/31/25, 11:38 AM

yes true

Ddaroth its a paper launch

Furkan Gözükara SECoursesOP•1/31/25, 11:38 AM

yes i watched

Furkan Gözükara SECoursesOP•1/31/25, 11:38 AM

so so shameless

Furkan Gözükara SECoursesOP•1/31/25, 11:38 AM

they produced 100,000 H100 for twitter in few weeks

Uumuthasanoglu Hi guys, which captioning model is the best to use with SUPIR upscaling? Also wh...

Furkan Gözükara SECoursesOP•1/31/25, 11:38 AM

llava best

Furkan Gözükara SECoursesOP•1/31/25, 11:39 AM

we have 2 config

Furkan Gözükara SECoursesOP•1/31/25, 11:39 AM

compare them

Furkan Gözükara SECoursesOP•1/31/25, 11:39 AM

look presets on interface

FFurkan Gözükara SECourses yes true

watersocks•1/31/25, 4:53 PM

ok awesome, any suggestion on why the model won't load? I've tried several times before but have never gotten it to work and always end up giving up, but I can't keep spending so much on credits for other paid services lol

Wwatersocks ok awesome, any suggestion on why the model won't load? I've tried several times...

AiInfluence•1/31/25, 5:22 PM

strange it doesn't work

AiInfluence•1/31/25, 5:23 PM

What's your root folder on swarm

SnicksterJay•1/31/25, 6:24 PM

Still doing research into hunyuan lora training?

MeLL•1/31/25, 7:39 PM

Can anyone help me with this error?

MMeLL Can anyone help me with this error?

AiInfluence•1/31/25, 9:19 PM

In the Models tab , find the burger icon and select edit meta data , from there set the architecture of the model to Flux Fill . For some reason Swarm doesn't do it automatically for different quantized versions

Wwatersocks ok awesome, any suggestion on why the model won't load? I've tried several times...

Furkan Gözükara SECoursesOP•1/31/25, 10:34 PM

maybe download was corrupted

SSnicksterJay Still doing research into hunyuan lora training?

Furkan Gözükara SECoursesOP•1/31/25, 10:34 PM

yes it is on book

MMeLL Can anyone help me with this error?

Furkan Gözükara SECoursesOP•1/31/25, 10:34 PM

try to put into stablediffusion folder

AAiInfluence In the Models tab , find the burger icon and select edit meta data , from there ...

Furkan Gözükara SECoursesOP•1/31/25, 10:35 PM

ah yes this is another possibility

Furkan Gözükara SECoursesOP•1/31/25, 10:35 PM

trying to fix paints undo scripts that we have

FFurkan Gözükara SECourses try to put into stablediffusion folder

MeLL•1/31/25, 10:50 PM

I have all this in a folder but changing the name of checkpoints to Stable-Diffusion doesn't help

AAiInfluence In the Models tab , find the burger icon and select edit meta data , from there ...

MeLL•1/31/25, 10:51 PM

dont work look
I have all this in a folder but changing the name of checkpoints to Stable-Diffusion doesn't help

MeLL•1/31/25, 10:55 PM

It's not the model that's the problem because it's not working with any...

MeLL•1/31/25, 10:56 PM

no model that I put through the patch..
and I don't have space to duplicate all the files

MMeLL dont work look I have all this in a folder but changing the name of checkpoint...

Furkan Gözükara SECoursesOP•1/31/25, 11:03 PM

did you edit metadata?

Furkan Gözükara SECoursesOP•1/31/25, 11:03 PM

if so please ask here : https://discord.gg/KfSd3ABg

FFurkan Gözükara SECourses did you edit metadata?

MeLL•1/31/25, 11:05 PM

model metadata you say?
Yes, I looked at what the boy said...
that was wrong

FFurkan Gözükara SECourses faster but i dont know if easier

sergio_gonzo•2/1/25, 2:10 AM

Dr Furkan, for kohya trinaing there is another flux dev model that could be used for training instead fo flux dev 1?

Rockerx•2/1/25, 8:25 AM

If I want to further finetune a lora or a model, what is the best approach ? Do i need the original dataset ? or can i give the model or lora new images without the old ones and continue further tuning ?

�

𝗡𝗶𝘁𝗶𝗻 𝗣𝗮𝗿𝗲𝘃𝗮•2/1/25, 12:05 PM

@Dr. Furkan Gözükara Is it possible to train a style and an object in the same Lora? You mention something about having two folders in the dataset preparation part and having a specific number of repeats for each of them.

rob•2/1/25, 2:43 PM

What are the fundamental differences between kohya, onetrainer and simpletuner? I've only used ai-toolkit (kohya) because there are many online resources, but I've seen information about SimpleTuner being better.. however, I am not sure why/what the differences could be

AiInfluence•2/1/25, 7:24 PM

AiInfluence•2/1/25, 7:24 PM

Some stuff I made with rank 128

Rockerx•2/2/25, 4:49 AM

@Dr. Furkan Gözükara can you suggest ?
If I want to further finetune a lora or a model, what is the best approach ? Do i need the original dataset ? or can i give the model or lora new images without the old ones and continue further tuning ?

mmercuri•2/2/25, 7:05 AM

Whats the recommended ui/ config to use flux with controlnet?

MMeLL Can anyone help me with this error?

Vitium•2/2/25, 8:30 AM

Are you, by any chance, using swarmUI through StabilityMatrix? I had a similar issue ( couldn't load any flux model whatsoever), and I highly suspect it's StabilityMatrix being wonky

faster but i dont know if easier

Similar Threads