Stable diffusion thread

[x1243x]

not bad.
i tried "a chinese female model seductive look fashion"

i think the result is more or less quite similar to sdxl.

but even sd1.5 can do it if you setup peturbed attention and self-attention guidance together (not including the text part. that one sd1.5 cannot do).


now which webui can run locally using GTX 4060?
i am still using sd forge coz i dont like comfyui.
swarm ui is also not very userfriendly in my opinon.

sd forge has stopped it's support though. that's why I moved to comfyui

 

[focus1974]

sd forge has stopped it's support though. that's why I moved to comfyui

i will keep using it as long as it is working. not upgrading
hope swarmui can evolve to be better. i want userfriendly but with controlnet type.

 

[BlackWing1977]

Finally followed some instructions to the tee posted on AMD Amuse thread... to disable the moderation... and it is perfect.

 

[doogyhatts]

Lumina has also released a new photorealistic model.
https://github.com/Alpha-VLLM/Lumina-mGPT

 

[doogyhatts]

Haven't tried writing my own custom node yet.

 

[doogyhatts]

People on reddit anyhow think the Flux.dev model can be used for commercial purposes.
Here is my understanding.

From the flux-dev license page,
“Non-Commercial Purpose” means any of the following uses, but only so far as you do not receive any direct or indirect payment arising from the use of the model or its output: ...

Just replace the capitalised words (Non-Commercial Purpose) with the entire definition.

Non-Commercial Use Only. You may only access, use, Distribute, or creative Derivatives of or the FLUX.1 [dev] Model or Derivatives for any of the following uses, but only so far as you do not receive any direct or indirect payment arising from the use of the model or its output.

You may use Output for any purpose (including for commercial purposes), except as expressly prohibited herein.

So it is actually prohibited if I re-read the above portion on Non-Commercial Use Only.

I am very certain only the outputs from the Pro version can be used commercially.
The Dev version is for development usages, such as an artist using the output-image for rotoscoping of a character in a specific pose. In other words, it is meant to increase an artist's productivity. And the output-images are not used directly in the final product.

See what the authors over at Black Forest Labs reply.

 

[BlackWing1977]

TS is perm banned ar???

 

[doogyhatts]

TS is perm banned ar???

I didn't notice either.
No wonder he was silent for so many days.

 

[doogyhatts]

I noticed my YT channel now has a dedicated Japanese fan, who seems to be on the intelligent side.
He will always reply where possible and he gave me some suggestions on what video to make.

Only cleared the first hurdle of 50 hours of watch time.

 

[BlackWing1977]

Just check ar.. this thread say KYM.. but did anyone got into trouble for posting cute AI zehzeh??

I know that posting lingerie-like outfit is sure kena.. but if normal clothing like a dress but revealing cleavage will kena or not??

 

[AZE]

New image model called Flux by the original team for Stable Diffusion.
12B parameters size. Need min 16gb vram. Recommended 24gb vram.
Midjourney quality.
HF demo.



I tried the HF demo, it couldn't do the same quality for the monsters that I had been generating using PixArt-Sigma.

Flux a bit too large for consumer grade HW. It can most probably run but training will be very hard to achieve.
One of the reasons other models rike the original SD series thrives ish due to possibility to train models with consumer grade HW or at least do minimally high quality finetunes.
Together with more open and permissive licensing, that forms the basis of the continued existence of the community.

 

[AZE]

Just check ar.. this thread say KYM.. but did anyone got into trouble for posting cute AI zehzeh??

I know that posting lingerie-like outfit is sure kena.. but if normal clothing like a dress but revealing cleavage will kena or not??

Here ish FFF grade forum.

 

[doogyhatts]

Flux a bit too large for consumer grade HW. It can most probably run but training will be very hard to achieve.
One of the reasons other models rike the original SD series thrives ish due to possibility to train models with consumer grade HW or at least do minimally high quality finetunes.
Together with more open and permissive licensing, that forms the basis of the continued existence of the community.

The results I got also don't look suitable from the HF demo for flux1.dev model.

I still got better looking monsters from Pixart-Sigma.
And I am getting plenty of watch time on the YT channel for those monsters with sound effects.
So no reason to seek out a new image model.

Flux1.dev licensing has contradictions. I treat it as strictly non-commercial.
So I am not going to waste my time on Flux.
Many people using Flux1.dev because they want to create naked AI women and sell those images.
The only way they could do that is to through the derivative models, from the goodwill of someone paying for access to a A100 gpu for training additional nsfw images.

 

[BlackWing1977]

The results I got also don't look suitable from the HF demo for flux1.dev model.

I still got better looking monsters from Pixart-Sigma.
And I am getting plenty of watch time on the YT channel for those monsters with sound effects.
So no reason to seek out a new image model.

Flux1.dev licensing has contradictions. I treat it as strictly non-commercial.
So I am not going to waste my time on Flux.
Many people using Flux1.dev because they want to create naked AI women and sell those images.
The only way they could do that is to through the derivative models, from the goodwill of someone paying for access to a A100 gpu for training additional nsfw images.

Naked AI women is probably the greatest reason why so many guys are into AI generative...

 

[BlackWing1977]

 

[jadedRomeo]

VR-Man reimagined -

PCK reimagined -

Obligatory waifu post -

How are u guys liking Flux so far? I find it abit more compliant than SD3.
I find I don't even need LORA for Flux.

 

[doogyhatts]

VR-Man reimagined -

PCK reimagined -

Obligatory waifu post -

How are u guys liking Flux so far? I find it abit more compliant than SD3.
I find I don't even need LORA for Flux.

Only tried the Dev version on their HF demo. Didn't try the Pro version.
Prompt adherence is good for people characters but couldn't get a cinematic artistic style.
Lack of diversity and doesn't work well with monsters.

The good anatomy from Flux won't matter because YT now has been sufficiently trained to disengage recommendation of AI-girls.
So all the generated images from Flux that have scantily clad, bikini, lingerie, sexy poses, cannot be monetised on YT.
Even fully clothed young female characters or young children can be implicated as well.

Old woman and old man still works though.

 

[ice888]

VR-Man reimagined -

PCK reimagined -

Obligatory waifu post -

How are u guys liking Flux so far? I find it abit more compliant than SD3.
I find I don't even need LORA for Flux.

What did u use for the img to vid?

 

[jadedRomeo]

What did u use for the img to vid?

Kling.
I prefer it over luma.
Haven't tried runway though.

 

[AikiBoy]

side track abit... is there a free good text to speech program or app