Software Engineering Courses (SECourses)•10mo ago

I think the latest DiffSynth-Studio fix for FP8 on WAN 2.1 14B Image-to-Video 480P has broken the FP

I think the latest DiffSynth-Studio fix for FP8 on WAN 2.1 14B Image-to-Video 480P has broken the FP16 quality too. I'm testing with previous commit before the FP8 fix

JJaviEM I think the latest DiffSynth-Studio fix for FP8 on WAN 2.1 14B Image-to-Video 48...

JaviEMOP•3/14/25, 2:38 PM

Confirmed

JaviEMOP•3/14/25, 2:38 PM

@Dr. Furkan Gözükara Please revert (or fix) the latest DiffSynth update. It breaks colours and quality in FP16, and in FP8 the video turns completely pink. I think it only affects WAN 2.1 14B Image-to-Video 480P model

PixelMonkey•3/14/25, 3:25 PM

@JaviEM i do noit have that iissue .. strange..
I do have a issue so running video to video

RuntimeError: The size of tensor a (36) must match the size of tensor b (13) at non-singleton dimension 2

PPixelMonkey @JaviEM i do noit have that iissue .. strange.. I do have a issue so running vi...

JaviEMOP•3/14/25, 3:48 PM

I reverted to previous DiffSynth commit and it's working fine again, so it's a problem from the latest commit

JJaviEM I reverted to previous DiffSynth commit and it's working fine again, so it's a p...

PixelMonkey•3/14/25, 4:11 PM

Ok that is weird .. for me it seems to work fine on rtx 3090

JJaviEM WAN 2.1 14B Image-to-Video 480P FP8 is broken

Furkan Gözükara SECourses•3/14/25, 4:38 PM

the offical repo authors made some fixes

Furkan Gözükara SECourses•3/14/25, 4:38 PM

i will move to them back

Furkan Gözükara SECourses•3/14/25, 4:38 PM

i hope it fixes

TTrumpYouAll @Dr. Furkan Gözükara I just come back and thank you for: 1. the Dreambooth trai...

Furkan Gözükara SECourses•3/14/25, 4:38 PM

1 : yes

Furkan Gözükara SECourses•3/14/25, 4:38 PM

2 awesome

JJaviEM And the negative prompt box is always empty when you start the app

Furkan Gözükara SECourses•3/14/25, 4:38 PM

will fix this too

PPixelMonkey @JaviEM i do noit have that iissue .. strange.. I do have a issue so running vi...

Furkan Gözükara SECourses•3/14/25, 4:39 PM

this looks like frame count of video to video

Furkan Gözükara SECourses•3/14/25, 5:10 PM

i am testing newest DiffSynth i hope they added lora properly as we did

blackhat•3/14/25, 5:18 PM

@Dr. Furkan Gözükara can you tell me the commands need to start the swarmui in the runpod bcz i installed and runned it last time with one command and now i have installed i want to run that only so i can use gui to do some generations

Bblackhat @Dr. Furkan Gözükara can you tell me the commands need to start the swarmui in t...

Furkan Gözükara SECourses•3/14/25, 5:42 PM

i have updated commands here https://www.patreon.com/posts/tutorial-instructions-links-public-post-106135985

Patreon

SwarmUI Master Tutorial - Use Stable Diffusion 3.5 Large and FLUX m...

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

Furkan Gözükara SECourses•3/14/25, 5:43 PM

and tutorial here

Furkan Gözükara SECourses•3/14/25, 5:43 PM

https://youtu.be/XFUZof6Skkw

YouTubeSECourses

How to Use SwarmUI & Stable Diffusion 3 on Cloud Services Kaggle (f...

In this video, I demonstrate how to install and use #SwarmUI on cloud services. If you lack a powerful GPU or wish to harness more GPU power, this video is essential. You'll learn how to install and utilize SwarmUI, one of the most powerful Generative AI interfaces, on Massed Compute, RunPod, and Kaggle (which offers free dual T4 GPU access for ...

blackhat•3/14/25, 5:43 PM

thanks a lot

Furkan Gözükara SECourses•3/14/25, 5:43 PM

i have to be afk few hours after that hopefully will fix wan 2.1 errors

smoke•3/14/25, 5:54 PM

@Dr. Furkan Gözükara Would you recommend to always use adetailer for faces (even if it’s a closer face)? Using forge.

And do you know if adetailer would possibly lower likeness of a character lora? Or if it's supposed to increase it

Ssmoke @Dr. Furkan Gözükara Would you recommend to always use adetailer for faces (even...

PixelMonkey•3/14/25, 6:07 PM

hey.. i used forge before as well and switched to SwarmUi and can highly recommend it. I use the facedetaile yolo there and it works amazing. There is no way to achieve that detail out of the box in my opinion, especially witzh multiple character images

PPixelMonkey hey.. i used forge before as well and switched to SwarmUi and can highly recomme...

smoke•3/14/25, 6:16 PM

Will have a look into that! Thanks a lot

Nrkozan•3/14/25, 6:20 PM

I'm looking for a ComfyUI workflow that generates output using a pose and reference image with LoRA.

Does anyone have such a workflow, or is there someone who can help as a freelancer?

xlads•3/14/25, 7:54 PM

Hey guys anyone know which tut shows how to create a preset?

blackhat•3/14/25, 8:11 PM

why SwarmUI on top if ComfyUi is working fine ? anyone can tell me ?

Furkan Gözükara SECourses•3/14/25, 9:43 PM

back working now on fixes

Ssmoke @Dr. Furkan Gözükara Would you recommend to always use adetailer for faces (even...

Furkan Gözükara SECourses•3/14/25, 10:12 PM

for closer shots it may not be necessary

Xxlads Hey guys anyone know which tut shows how to create a preset?

Furkan Gözükara SECourses•3/14/25, 10:12 PM

presets in which app swarm?

Furkan Gözükara SECourses•3/14/25, 10:13 PM

this one : https://youtu.be/HKX8_F1Er_w

YouTubeSECourses

Zero to Hero Stable Diffusion 3 Tutorial with Amazing SwarmUI SD We...

Do not skip any part of this tutorial to master how to use Stable Diffusion 3 (SD3) with the most advanced generative AI open source APP SwarmUI. Automatic1111 SD Web UI or Fooocus are not supporting the #SD3 yet. Therefore, I am starting to make tutorials for SwarmUI as well. #StableSwarmUI is officially developed by the StabilityAI and your m...

Bblackhat why SwarmUI on top if ComfyUi is working fine ? anyone can tell me ?

Furkan Gözükara SECourses•3/14/25, 10:13 PM

what do you mean?

FFurkan Gözükara SECourses what do you mean?

blackhat•3/14/25, 10:15 PM

i created a model with the flux dev and it is working fine it took only 10 mintues and giving me good outputs and i also tried with the way you explained in the tutorials with massive computation so what is diff , i am doing it with the flux and the 10 images of data set and getting same quality of the iamges as expected from your model and mine are same .

what difference will be created from this

blackhat•3/14/25, 10:17 PM

i am talking in the context of the person flow as you stated

blackhat•3/14/25, 10:19 PM

i want variety in expression and consistency in the body of the person but i guess as you stated that can be gathered by dataset if we include them in that

so i want to improve my person flow and the generations to get eveyhthing with prompts instead of big dataset

Bblackhat i want variety in expression and consistency in the body of the person but i gue...

Furkan Gözükara SECourses•3/14/25, 10:32 PM

you need better dataset

Furkan Gözükara SECourses•3/14/25, 10:32 PM

if you want better expressions

Furkan Gözükara SECourses•3/14/25, 11:19 PM

wow new teacache really speeds up

Furkan Gözükara SECourses•3/14/25, 11:23 PM

new teacache

Furkan Gözükara SECourses•3/14/25, 11:23 PM