Stable diffusion thread

[doogyhatts]

Hmm... I think still like quite complex to setup.... that is why I until now still yet to go try ComfyUI... and really dun have the interest to keep trying all these also..

But I do think eventually will be so common until a more refined solution will be available also.

Wan2GP should be easy to setup on Windows, also supports AMD GPUs.
https://github.com/deepbeepmeep/Wan2GP

 

[doogyhatts]

Another image model from Alibaba, which focuses on complex text rendering.

 

[doogyhatts]

 

[doogyhatts]

This is the new way to do image and video generation. Just use natural language and tags.
I managed to get to do some beta testing over the last four days.

The user guides have the details.
https://app.klingai.com/global/quickstart/klingai-image-o1-user-guide
https://app.klingai.com/global/quickstart/klingai-video-o1-user-guide

Release notes shows Runway Aleph as closest competitor.
https://app.klingai.com/global/release-notes/vaxrndo66h?type=dialog

 

[doogyhatts]

A cooler way to explain by FreePik.

 

[doogyhatts]

@AikiBoy
Kling now is very different. It can now do both image and video editing.

 

[TimsTom]

A cooler way to explain by FreePik.

Is this foc to use or need pay?

 

[doogyhatts]

Is this foc to use or need pay?

Need to pay one.
But Pro-plan subscribers and above (before 15th dec 2025), gets to use the Image-O1 model for free during off-peak durations for 1 year.

 

[AikiBoy]

@AikiBoy
Kling now is very different. It can now do both image and video editing.

Nice.

Recent i see many tiktok Ai by local ppl.. duno who are they, which AI they use

 

[doogyhatts]

SteadyDancer motion capture research.
https://mcg-nju.github.io/steadydancer-web/

Now available to use on Comfy and Wan2GP.

 

[doogyhatts]

@AikiBoy Nano-banana-pro, Midjourney-v7, and GPT-image-1 are now on Kling Lab.

 

[HushiePuppies]

now got z-image-turbo, everything else pales in comparison.

 

[focus1974]

now got z-image-turbo, everything else pales in comparison.

NSFW can?

OMFG... CAN!!

 

[doogyhatts]

NSFW can?
OMFG... CAN!!

But we cannot monetise nsfw images/videos, unlike others in Japan.
US payment processors also don't allow such things on their transactions.

Only useful for self-entertainment and fix blurry human anatomy as a first pass.

 

[doogyhatts]

now got z-image-turbo, everything else pales in comparison.

Z-image alone not enough, still need Wan 2.2 Low-Noise for latent upscaling and SeedVR2 for pixel upscaling.

 

[focus1974]

But we cannot monetise nsfw images/videos, unlike others in Japan.
US payment processors also don't allow such things on their transactions.

Only useful for self-entertainment and fix blurry human anatomy as a first pass.

really is for self-entertainment. lol

 

[doogyhatts]

2x2 from NBP is a better choice than 3x3.

 

[doogyhatts]

Ubisoft Open-Sources the CHORD Model and ComfyUI Nodes for End-to-End PBR Material Generation
https://blog.comfy.org/p/ubisoft-open-sources-the-chord-model​

“Considering the multi-stage nature of our prototype, ComfyUI provides us with an efficient framework to build integrated workflows doing texture image synthesis, material estimation and material upscaling. This also enables us to leverage state-of-the-art generative models and the powerful features of ComfyUI that provide fine-grain control to creators with ControlNets, image guidance, inpainting, and countless other options.”
— Ubisoft La Forge Blog