GiraffeInTheMatrix

joined 8 months ago
sorted by: new top controversial old
2
Whatever change happened to the image model just now screwed up the prompting. (lemmy.world)
 

It's not doing parts of the prompt I put in anymore, it's just ignoring it or making it look worse then before.

The old version is better in c/[email protected]
[–] [email protected] -1 points 1 week ago (1 children)

Why is everyone trying to dislike bomb this.

The old version is better in c/[email protected]
[–] [email protected] 1 points 2 weeks ago

The furry artstyles get a bit screwy. They keep trying do the style not set and do a similar but totally different one.

Will e621 furry tags still be able to be imported from plugins as pointers for image generation? or is the limitations and complexity of Flux.1 not allow it anymore thus making it impossible? in c/[email protected]
[–] [email protected] 2 points 2 weeks ago* (last edited 2 weeks ago)

@[email protected] can I get insight on this question.

Can you give us an insight on progress of the update with the new model with notifications for the community when you small updates and changes with the new Model like what's planned next, what your currently working on, what things have been added, etc. Like a Devlog for us. And for one more thing what is going to happen to the old model?

It would be really helpful for us folks who would like to know what's going on, and it would stop the rift of the people who loved the old model and the ones who adore the new model as well as calming the community.

“I’m Not Asking for Magic — Just Consistency, Like Before” bring the old model back please in c/[email protected]
[–] [email protected] 2 points 2 weeks ago

I support bringing back weight prompting but for flux.1 it's impossible and ignores it completely. Plus the Dev won't change it back. He already made up his mind about it, no matter how much some parts of the community want it back its dwarfed by all the praise by people who absolutely adore the new model.

“I’m Not Asking for Magic — Just Consistency, Like Before” bring the old model back please in c/[email protected]
[–] [email protected] 4 points 2 weeks ago (2 children)

Bro chill. I get that the people who want the old model back are getting on your nerves but jeez

4
Will e621 furry tags still be able to be imported from plugins as pointers for image generation? or is the limitations and complexity of Flux.1 not allow it anymore thus making it impossible? (lemmy.world)
Users rising thanks to ai art update :) in c/[email protected]
[–] [email protected] 2 points 2 weeks ago

Well I get that, a bunch of people were upset and left. Then new people filled in the losses.

@[email protected] will the dynamicimport plugin work for the new model sooner or later? It isn't working well. in c/[email protected]
[–] [email protected] 1 points 3 weeks ago (1 children)

I was trying to get it to do more artstyles that were from the E621 import

1
@[email protected] will the dynamicimport plugin work for the new model sooner or later? It isn't working well. (lemmy.world)
We'll probably be stuck with Perchance's Flux.1 model. The dev won't add SDXL or any type of SD model as a secondary. This is my current conclusion of the model controversy. in c/[email protected]
[–] [email protected] 2 points 3 weeks ago (1 children)

It's just a thing to point out to people that still want the old model that is pretty much pointless to ask for the old model back and stuff.

3
We'll probably be stuck with Perchance's Flux.1 model. The dev won't add SDXL or any type of SD model as a secondary. This is my current conclusion of the model controversy. (lemmy.world)
 

I miss the old model, but we can't do anything about it anymore. I think dev made up his mind on having Flux.1 be permanent for all of Perchance AI image generators.

The people who adore the new model have won over the Developer of Perchance. It's pointless of asking for the old model or requesting SDXL as a secondary image model. I've mostly given up/ forever paused my image generator like some others have with their own generators.

It's not going to change.

Best alternatives similar to the old version? in c/[email protected]
[–] [email protected] 1 points 3 weeks ago* (last edited 3 weeks ago)

There is one called Frosting.Ai which I switched too. It's mostly used for adult 18+ content but you can generate other things on there aswell.

It's not very good sometimes but it uses the same format as SD and you gotta pay for Weight Prompting which is a downside. But some people can get good generators on the free versions. It has a drawback of limited prompting but it has unlimited number of image generations without those stupid credit stuffs

FIXED - political prompt working again in c/[email protected]
[–] [email protected] 2 points 3 weeks ago

Oh yeah that happens a lot.

3
Weird thing, the new Image Model can make pro nazi propaganda. I generated a poster thing as a joke to see if it would do it and it did it. (lemmy.world)
 

Is this a concern?

2
There's a bunch of toxicity now. The people who like the new model are just making fun of the people who dislike it. And Vice versa. I apologize. I was just frustrated. (lemmy.world)
submitted 1 month ago* (last edited 1 month ago) by [email protected] to c/[email protected]
2 comments fedilink
 

At this point if the toxicity keeps going and it gets worse because of this divide.I apologize for being out of line I was frustrated and stuff and I senserly apologize.

I plan on hosting the previous text-to-image model on my own. But I need to know the EXACT model of Stable Diffusion. in c/[email protected]
[–] [email protected] 1 points 1 month ago (1 children)

From what i looked up its a mix of models. Flux Ai, SDXL, SDXL Lightning, and SDXL Flash.

I HATE NEW MODEL in c/[email protected]
[–] [email protected] 4 points 1 month ago

Let's just hope they switch to something better and more familiar then Flux.

5
To the developer of Perchance.. I have a suggestion as well as a request from me and some of my fellow users. May you consider switching from Flux to the SDXL model? Please consider this! <3 (lemmy.world)
submitted 1 month ago* (last edited 1 month ago) by [email protected] to c/[email protected]
2 comments fedilink
 

I’d like to offer a suggestion that's a request from me and the rest of the user base that adopted SD 1.5 as our main core of our generators. We request that you please consider switching from the Flux Model to the SDXL model for Perchance's new model of the text to image generation process.

We think SDXL would likely be a much better fit for Perchance's AI image needs and your website's user base.

SDXL can make a familiar but mu h improved output of images, but also it easier for new users to quickly learn and understand prompting easily, and helping especially those already familiar with the original prompting of the old model of SD 1.5 without having to completely change their precious prompts that they worked hard to perfect.

According to a friend of mine SDXL is also is much less expensive then Flux and easier for the community to use. The switch could reduce backend complexity while offering better image quality andfaster generation times. This will give the users something they can use and keep using their original prompts rather then making it into a huge long essay of confusion and words.

Thank you!

0
(Feature Request) Adding AI models for AI image Generation (lemmy.world)
 

I've always wanted to add a Ai Image Model from Civitai. I'm wondering when the option will be added in a later update. I'm sure others would enjoy this feature.

view more: next ›