Support NeoGAF

[RedC]

 

[RedC]

 

[RedC]

 

[RedC]

 

[RedC]

 

[Solarstrike]

 

[RJMacready73]

Those are awesome, what sorta prompts did you use to get that?

 

[Solarstrike]

Those are awesome, what sorta prompts did you use to get that?

Thanks RJ!

First image: 8k detailed Wizard wearing a detailed cloak standing on a cliff overlooking a detailed city, detailed landscape, detailed castle, detailed rock geodes, detailed flora

Second image: 8k detailed warrior wearing worn metal and leather armor, battle-worn, holding an ancient flag, walking on a dirt path in the rain, movement action scene, detailed landscape, detailed worn flag

Third image: 8k detailed whimsical scene, magical scene, astronaut walking up a hill on a strange planet, detailed and serene landscape, strange flora, strange rocks

Fourth image: 8k detailed closeup of a creature peering (or peeking) through a hole in a barn, detailed woodgrain, detailed and accurate eyes and no mouth, cryptid Dover Demon (or Jersey Devil). The name is Dover Demon on my DeviantArt page (PenitentMachine)

 

[ResurrectedContrarian]

if Disney never purchased Star Wars

 

[-Minsc-]

"cybersix animated"

Not Cybersix. I'm curious to know what it's drawing from. It did know enough to go with black leather, cape and black hair. The red rose is out of place. It must have somehow morphed the red lining of the cape into a rose

 

[Boralf]

 

[ManaByte]

Noticed this morning prompts are extremely limited and output has taken a massive step backwards.

Must be due to that AI Hitler photo someone made and is going viral. They're putting the hammer down.

 

[Boralf]

I don't see any difference, But maybe that's because I'm tunnelvisioning on the same prompt the whole time.

Btw, what are these weird white (with a bit of blue/purple) spots sometimes? You see one here in the middle for example on the right side of the stairs. There's also on in the middle top in the first picture in my previous post.

 

[ManaByte]

I don't see any difference, But maybe that's because I'm tunnelvisioning on the same prompt the whole time.

Been getting a lot more blocked prompts this morning. They seem to be tightening the censorship A LOT more. They're scared due what some dumbass did with Hitler.

 

[RedC]

I didn't care about downloading the pictures I generated via Dall-E 3 and was cotent with simply using the Dall-E 3 links to upload the images.

When the images started disappearing i still didn't care.

However going through this thread and seeing most of my posts with no data somehow bothered me so I finally downloaded the remaining images of my posts on the last 3 day pages and uploaded them to an image host and reuploaded them individually on each post and it was a pain in the ass.

Now it has me regretting not doing it from the jump and it could be a time capsule on the evolution and censorship I was able to generate using Dall-E 3 over an extended period of time.

 

[Solarstrike]

.

 

[RedC]

 

[Solarstrike]

.

 

[violence]

Anyone tried Grok image generator? It has fewer guard rails.

 

[RedC]

Anyone tried Grok image generator? It has fewer guard rails.

Is there a free version?

 

[ResurrectedContrarian]

Grok uses Flux, new model from the real team behind the original Stable Diffusion models.

The "pro" version that Grok is probably using isn't free, but you can access it for basically pennies via sites like FAL https://fal.ai/models/fal-ai/flux-pro

Or you can download and run (or train) the still-impressive Dev version of the model on your own hardware without limits. And there are plenty of community-trained modifications for it.

 

[Solarstrike]

"Routine Exam"

 

[RedC]

 

[Project EVE]

Any reason to keep using Dall-E 3 when Flux.1 is much less censored?

 

[Solarstrike]

Any reason to keep using Dall-E 3 when Flux.1 is much less censored?

Can use whatever you want, I imagine! But maybe EviLore can clarify. Thank you

 

[Trogdor1123]

Any reason to keep using Dall-E 3 when Flux.1 is much less censored?

What’s is the difference? Never used flux.

 

[Solarstrike]

.

 

[EverydayBeast]

 

[Solarstrike]

.

 

[RedC]

 

[RedC]

 

[RedC]

 

[RedC]

 

[Lord Panda]

Anyone tried Grok image generator? It has fewer guard rails.

Yeah I'm using Flux.1 (dev). Here's some of mine. (Disclaimer: all images are for fun and initially generated to test the extent of the training and censorship imposed if any. They are and not in any way political statements etc)

 

[Solarstrike]

"Morning Hymn"

 

[tkscz]

Here's something fun you can do, but may take time to render.
Go to Luma Dream Machine's AI Video creator, upload your image from Dall-E3 and tell it to pan the camera in a specific direction with it's camera commands (move left, right, pull in, push out)

It's not always perfect but it can come up with some nice results


Another example:

 

[RedC]

 

[RedC]

 

[RedC]

 

[Solarstrike]

.

 

[RedC]

 

[Solarstrike]

 

[RedC]

 

[Hrk69]

My megalophobia is kicking in

 

[Solarstrike]

.

 

[RedC]

 

[RedC]

 

[Radioskugga]

 

[tkscz]

 

[RedC]