Stable diffusion thread

[doogyhatts]

@AikiBoy
Kling 2.1 master T2V model now costs 40% less credits for subscribers.

 

[AikiBoy]

Hmmm too much competition

 

[x1243x]

I figured out a few other mistakes I made.

There is a large group of zombie subscribers in my channel.
To avoid this group, what is recommended in the post, is to release the video in the middle of the night (1am), when they are sleeping.
This also helps to avoid the group of zombie subscribers.

The main issue is actually on comments.
The more the video gets in its initial first 24 hours, the better it would perform in the week.
This also helps to push past the group of zombie subscribers.

One way to get comments is to ask what time and where the viewers are watching from.
The Sleepless Historian does exactly this for every video.

1am sg time?

 

[doogyhatts]

1am sg time?

Yes.

But the timing alone is not sufficient as it still needs many comments.
Three other possibilities to get comments:
1) comments on the lore/story for an upcoming game. (BM: Zhong Kui)
2) comments on a past game played before long ago. (WoW)
3) comments on a newly discovered SCP story.

 

[x1243x]

Yes.

But the timing alone is not sufficient as it still needs many comments.
Three other possibilities to get comments:
1) comments on the lore/story for an upcoming game. (BM: Zhong Kui)
2) comments on a past game played before long ago. (WoW)
3) comments on a newly discovered SCP story.

interesting. ok I'll give it a try

 

[doogyhatts]

Text-to-audio and video-to-audio.

 

[psyger-zero]

Text-to-audio and video-to-audio.

will zheng ever get unbanned?

 

[doogyhatts]

will zheng ever get unbanned?

Probably not.

 

[glarerder]

When ish moi ai gf coming...

 

[doogyhatts]

Fast action transition styles.

 

[x1243x]

will zheng ever get unbanned?

how/why did he get banned?

 

[doogyhatts]

>8 mins video, 29k views in 9 days, with a lot of comments.
Warcraft AI-content still has less competition compared to BM:ZK and SCP.

 

[doogyhatts]

Two different programs from KlingAI.

 

[doogyhatts]

ByteDance has updated OmniHuman to v1.5.
https://omnihuman-lab.github.io/v1_5/

 

[AikiBoy]

Google nano banana sibei powerful...
Better than chatgpt..

The result is scary. I upload 1 pic... i can change pose without losing consistency

 

[doogyhatts]

Bytedance releases a new LoRA for style transfer.

 

[AikiBoy]

Hailuo seem so expensive vs kling

 

[doogyhatts]

Hailuo seem so expensive vs kling

But Kling now only works when using start-end frames for Pro 2.1 model.
A lot of weird artifacts recently when using the master model.
I think they have changed something drastic that lowers the quality.

Despite that, they still gave me Ultra plan for free.

 

[doogyhatts]

 

[AikiBoy]

I think to use hailuo but at lowest plan cannot use some feature. Or generate 10s video.

Nano banana makes many thing easier...