Stable diffusion thread

[doogyhatts]

All paid services

Yes unfortunately.
Since it is only image-to-video, so it is best to prepare all the images first using the Standard plan before jumping into the Pro plan.
Because the video generation allows you to queue up to 40 jobs, while only two is processing in relaxed mode.

 

[doogyhatts]

Japanese bookstores move fast in AI.

 

[NintendoSwitch]

Japanese bookstores move fast in AI.

So when are the Japanese documentaries made by AI cumming?

 

[AikiBoy]

Yes unfortunately.
Since it is only image-to-video, so it is best to prepare all the images first using the Standard plan before jumping into the Pro plan.
Because the video generation allows you to queue up to 40 jobs, while only two is processing in relaxed mode.

Lol i feel guilty not uploading any video. Ideas run dry.

 

[AikiBoy]

Japanese bookstores move fast in AI.

Use skillsfuture to learn python now

 

[xcodes]

anyone knows what's the PC system requirements to run stable diffusion?

 

[doogyhatts]

anyone knows what's the PC system requirements to run stable diffusion?

For SD3.5, minimum vram required for GPU is 8gb.
For Flux, better to have at least 16gb vram.

 

[xcodes]

For SD3.5, minimum vram required for GPU is 8gb.
For Flux, better to have at least 16gb vram.

thanks bro ...

 

[focus1974]

For SD3.5, minimum vram required for GPU is 8gb.
For Flux, better to have at least 16gb vram.

sd3.5 is better than sd1.5?
Now i see the sd1.5 details are super realistic liao.

i got a 4060 8gb, so generate img will take forever or not?
now 1024x1024 img can be done <30seconds for sd 1.5 with controlnets.

 

[doogyhatts]

sd3.5 is better than sd1.5?
Now i see the sd1.5 details are super realistic liao.

i got a 4060 8gb, so generate img will take forever or not?
now 1024x1024 img can be done <30seconds for sd 1.5 with controlnets.

If you combine SD3.5 large with this lora, you can get Midjourney quality images.
https://huggingface.co/AIImageStudio/MidjourneyStyleRealisticFilmLoRA_SD3.5L

4060 8gb is fine, but the output images in the above site are of 832x1216.

fyi, @xcodes .

 

[AikiBoy]

So sad. No more free credit on hailuo

 

[doogyhatts]

So sad. No more free credit on hailuo

Do realistic WoW animations, won't go wrong.

 

[AikiBoy]

Do realistic WoW animations, won't go wrong.

I pretty those are trained AI for wow content.. random ai will prompt till u go crazy

 

[doogyhatts]

The Scarlet Monastery video now has 11k views in 2 days!
The clean style of its generated images is very similar to the outputs from Imagen 4.

 

[doogyhatts]

@AikiBoy
For characters, use the images from the WoW game.
Then use qwen chat or Google AI studio to describe the images.
Edit the prompts for photorealistic style.
After that, use Dreamina to create the images.

 

[AikiBoy]

https://vt.tiktok.com/ZSBUQVhLF/

I wonder what to uae to make this

 

[doogyhatts]

https://vt.tiktok.com/ZSBUQVhLF/

I wonder what to uae to make this

Should be Hailuo-02 model.

 

[AikiBoy]

Should be Hailuo-02 model.

Ahah but i kling user.

I found a fighting AI.. foto

 

[AikiBoy]

https://vt.tiktok.com/ZSByjbass/

This one say VEO3

 

[doogyhatts]

https://vt.tiktok.com/ZSByjbass/

This one say VEO3

Hmm... that means the previous one should be veo3 instead of hailuo-02.