Software Engineering Courses (SECourses)•10mo ago

flux dev put into diffusion_models folder

Furkan Gözükara SECoursesOP•2/26/25, 10:55 AM

yolo format : <segment:yolo-face_yolov9c.pt-1,0.7,0.5>photograph of ohwx man

FFurkan Gözükara SECourses flux dev put into diffusion_models folder

TrumpYouAll•2/26/25, 10:59 AM

@Dr. Furkan Gözükara thanks for your prompt reply, yes, I did use your installer from Patreon
"flux dev put into diffusion_models folder", may I know if my Dreambooth-trained models are put inside this folder, called 'diffusion_models' ? Is this folder inside 'Models' folder ?

Thanks in advance !

TTrumpYouAll @Dr. Furkan Gözükara thanks for your prompt reply, yes, I did use your installer...

Furkan Gözükara SECoursesOP•2/26/25, 11:36 AM

yes

Furkan Gözükara SECoursesOP•2/26/25, 11:36 AM

inside models

Furkan Gözükara SECoursesOP•2/26/25, 11:36 AM

diffusion_models

TrumpYouAll•2/26/25, 11:41 AM

@Dr. Furkan Gözükara as to the error message above as shown in the screenshot, would I be able to solve it by re-installing everything back again ?

TTrumpYouAll @Dr. Furkan Gözükara as to the error message above as shown in the screenshot, w...

Furkan Gözükara SECoursesOP•2/26/25, 11:42 AM

i would say yes

Furkan Gözükara SECoursesOP•2/26/25, 11:42 AM

you can also show debug logs

FFurkan Gözükara SECourses you can also show debug logs

TrumpYouAll•2/26/25, 11:52 AM

I would definitely give it a go again and see if this will solve the issue.
Thanks again for ur prompt reply @Dr. Furkan Gözükara

Furkan Gözükara SECoursesOP•2/26/25, 12:03 PM

i am adding custom resolutions to wan 2.1 app right now

Furkan Gözükara SECoursesOP•2/26/25, 12:25 PM

Furkan Gözükara SECoursesOP•2/26/25, 12:33 PM

new intrerface

Furkan Gözükara SECoursesOP•2/26/25, 12:33 PM

Furkan Gözükara SECoursesOP•2/26/25, 12:36 PM

Furkan Gözükara SECoursesOP•2/26/25, 12:36 PM

wow tiled made huge impact on speed

FFurkan Gözükara SECourses new intrerface

AiInfluence•2/26/25, 12:53 PM

any new video examples ?

daroth•2/26/25, 12:56 PM

damn a 4090 and 64GB Ram is not enough for local Wan 14B

Ddaroth damn a 4090 and 64GB Ram is not enough for local Wan 14B

Cyph3r•2/26/25, 1:01 PM

need 5090 even for the higher quantized or fp8 versions. need rtx 8090 for the full versions of the 14B Model lol

Ddaroth damn a 4090 and 64GB Ram is not enough for local Wan 14B

Timson•2/26/25, 1:09 PM

you may try to boot your machine in headless Linux OS to squeze it

Timson•2/26/25, 1:09 PM

windows gui take some ram and vram

daroth•2/26/25, 1:10 PM

the 2nd most powerful consumer gpu in existence and its not enough. existence is pain

Timson•2/26/25, 1:10 PM

this model were never ment for consumer grade GPU

Ddaroth damn a 4090 and 64GB Ram is not enough for local Wan 14B

Furkan Gözükara SECoursesOP•2/26/25, 1:11 PM

i generated on 96 gb 3090 ti

Furkan Gözükara SECoursesOP•2/26/25, 1:11 PM

takes very long

Furkan Gözükara SECoursesOP•2/26/25, 1:11 PM

therre is fp8 option can you test it?

Furkan Gözükara SECoursesOP•2/26/25, 1:11 PM

on my pc it fails since i have dual gpu

FFurkan Gözükara SECourses therre is fp8 option can you test it?

daroth•2/26/25, 1:14 PM

I see two quants on HF, which one you mean?

Ddaroth I see two quants on HF, which one you mean?

Furkan Gözükara SECoursesOP•2/26/25, 1:20 PM

no i mean this

daroth•2/26/25, 1:20 PM

Furkan Gözükara SECoursesOP•2/26/25, 1:22 PM

about to update for v3

Original message was deleted

Furkan Gözükara SECoursesOP•2/26/25, 1:25 PM

download v3 zip file please

Furkan Gözükara SECoursesOP•2/26/25, 1:25 PM

and

Furkan Gözükara SECoursesOP•2/26/25, 1:26 PM

use Windows_Install.bat

Furkan Gözükara SECoursesOP•2/26/25, 1:26 PM

v3 zip file here : https://www.patreon.com/posts/123105403

Furkan Gözükara SECoursesOP•2/26/25, 1:26 PM

delete wan v2 folder

Furkan Gözükara SECoursesOP•2/26/25, 1:33 PM

app is now v3 : https://www.patreon.com/posts/123105403

Patreon

Wan 2.1 Ultra Advanced Gradio APP for - Works as low as 4GB VRAM - ...

Get more from SECourses: FLUX, Tutorials, Guides, Resources, Training, Scripts on Patreon

Furkan Gözükara SECoursesOP•2/26/25, 2:33 PM

torch vision arrived amazing

Furkan Gözükara SECoursesOP•2/26/25, 2:33 PM

now all i have to do is compile flash attention

Jean-lu•2/26/25, 2:41 PM

Hi guys Im new here so please let me know if this question doesn't go here. But Im trying to train lora's using fluxgym I have a couple of questions / issues -

My training datatset is of kids (age 5-13) with max 15-20 photos. But even with an A100 40GB it takes around 2hrs to train. Any fix for that? I love how fal.ai does the fast training Im okay with even an hour but 2-3 gets tiresome
I tried with increasing batch size for perfomance but seems like they kinda ruin the quality. Any advice to how to maintain quality with perfomance? captions yay or nay? size 1024 or 512?

FFurkan Gözükara SECourses torch vision arrived amazing

1nfern0•2/26/25, 2:46 PM

does that mean you can stretch the 5090's legs soon. I haven't been able to test the app yet with work getting in the way.

JJean-lu Hi guys Im new here so please let me know if this question doesn't go here. But ...

Furkan Gözükara SECoursesOP•2/26/25, 2:47 PM

it is quality vs time

11nfern0 does that mean you can stretch the 5090's legs soon. I haven't been able to te...

Furkan Gözükara SECoursesOP•2/26/25, 2:47 PM

i am trying to compile flash attention noıw

FFurkan Gözükara SECourses ye

1nfern0•2/26/25, 2:47 PM

ohh wow, things are starting to move along nicely

11nfern0 ohh wow, things are starting to move along nicely

Furkan Gözükara SECoursesOP•2/26/25, 2:48 PM

yep

Furkan Gözükara SECoursesOP•2/26/25, 2:55 PM

Furkan Gözükara SECoursesOP•2/26/25, 2:55 PM

flash attention compile lets see i hope works

FFurkan Gözükara SECourses therre is fp8 option can you test it?

computersheep•2/26/25, 3:37 PM

FP8 will be extra slow on any gpu before the nvidia Lovelace generation
like A6000, 3090, A100 dont have the chips needed to perform well with FP8

L40, 4090, h100 would be good cards for FP8

just incase anyone cares

FFurkan Gözükara SECourses it is quality vs time

Jean-lu•2/26/25, 3:37 PM

Any idea how fal does it? the speed is amazing & quality is almost perfect too. Are they finetuning layers or just good GPUs

Ccomputersheep FP8 will be extra slow on any gpu before the nvidia Lovelace generation like A6...

Furkan Gözükara SECoursesOP•2/26/25, 3:44 PM

nice info thanks

flux dev put into diffusion_models folder

Similar Threads

Similar Threads

Similar Threads