Your AI-generated Art

AlphaJohn said:
Gothic painting of a castle
Click to expand...

That's a church, isn't it?


AlphaJohn said:
1700720624157.png

Victorian-era painting with office people working on computers
Click to expand...

Ah, so that's how the colonisers got ahead ...
 
DapperDanMan said:
That's a church, isn't it?
Click to expand...

I'm only posting the ones that kinda worked... for every one I post there are loads that failed badly ie:

1700722832118.png
Create a photo realistic image of a cat in a spacesuit outside a space station fixing a solar panel.

Granted most of my prompts are way out there, in order to examine its understanding.
 
MavChat said:
image.png

Victorian-era painting with office people working on computers - 0n SDXL - check the size of the keyboard!!
Click to expand...

Yeah so far I reckon we still have serious flaws in AI's understanding of the real world items. I can find errors in every pic generated example

"A raw photograph capturing a sustainable treehouse community nestled in a dense rainforest. The view is centered and taken from afar, showcasing the impressive and enormous treehouses on a sunny day, with delicate rays of light filtering through the hot, humid air. The perspective is from the viewer's point of view, standing on one of the treehouses, leaning on a wooden railing. The viewer's hands are visible on the railing, adding to the immersive experience. The surrounding jungle is lush and vibrant, with the treehouses blending seamlessly into the environment."

1700725918936.png1700726039400.png
Ke?

Ok it did give me one of what I asked for:
1700726097864.png

Dont get me wrong, its a big leap from when I last played on Dall-E 2 at the beginning of the year. But there is load of room left for improvement.
 
I like looking at the images posted here and try and replicate it using the provided prompts or prompts from what I see on SDXL... Results does vary as SDXL tends to only read prompts it wants to.
 
MavChat said:
I like looking at the images posted here and try and replicate it using the provided prompts or prompts from what I see on SDXL... Results does vary as SDXL tends to only read prompts it wants to.
Click to expand...

Ditto

Thats why I normally give part or the whole prompt for others to play with in case they looking for something similar.

The ChatGPT ones I normally only post part cause that dam thing can write books for a single image.
 
image.png

A faded photo of a forgotten greenhouse where plants took over, in the centre, in colour, is a woman in a gothic dress holding a single red rose
 
AlphaJohn said:
Ditto

Thats why I normally give part or the whole prompt for others to play with in case they looking for something similar.

The ChatGPT ones I normally only post part cause that dam thing can write books for a single image.
Click to expand...
SDXL does well with basic, to the point, prompts without feelings etc
 
You must log in or register to reply here.
Back
Top
Sign up to the MyBroadband newsletter
X