The art and photography thread (includes AI generated art)

ZedClampet

Community Contributor
Post your photographs and your art here to share with everyone. So far as I know, there are only two people here who can actually draw/paint, so we're including AI generated art. It does take a little bit of practice and skill to create good AI art, but of course we will all bow to the actual artists if any want to share their work.
 

ZedClampet

Community Contributor
I would like to get better at photography. I've got a good camera with all the necessary bells and whistles. Maybe when it's not way below freezing I'll contemplate taking my camera somewhere.

My wife, who you would never, ever suspect of being good at it is actually fantastic and takes amazing photos, so I usually just defer to her.

I used to do a lot of digital art, but I wasn't great at it, so I've switched to AI.
 
I bought a dslr about 20 years ago. It’s now worth about 5% of what I paid for it.

I tried to get good at photography but it never really clicked for me. I joined deviant art about the same time as a now quite famous photographer (for da - http://severinkoller.com/photography/ ) and he commented on my work but his work skyrocketed in popularity, and mine didn’t. Not that I had any of the talent he had. I just couldn’t understand the basics or get any inspiration.

Edit: what AI came up with in 60 seconds:

View: https://imgur.com/a/c5WzJvO
 
Last edited:
  • Like
Reactions: Frindis
This is a drawing I made when I was 10. It depicts my paradise:

alJKk06.jpeg


So far as I know, there are only two people here who can actually draw/paint

So who's the other one?
 

ZedClampet

Community Contributor
This is a drawing I made when I was 10. It depicts my paradise:

alJKk06.jpeg




So who's the other one?
I couldn't decide whether to go with the love or the laughing emoji on your post. I do love the picture.

I'm probably wrong, but for some reason I thought dxchase and johnway were the artists, but now you are the third

Somewhere on the site is an NFT of a boat I tried to sell to Brian.
 
  • Like
  • Love
Reactions: Frindis and Pifanjr
So I followed a guide to get Flux working in Pinokio. Took a good 20 minutes to install everything. Finally, Gradio shows up and lets me input a prompt. I decided to do something simple, so I inputted “two guinea pigs frolicking in a flowery meadow, cute pastel art style”. Pinokio then proceeded to attempt to download about 30gb of files, and after 10 minutes the image was still not generated. I terminated the request.

Is that normal for each prompt, or is it like a one time thing? I seen in the command prompt something about developer mode needing to be turned on so Hugging Face can more efficiently store cache, so I did that and restarted my PC. I then tried to make an image again and it still happened.

I have a RTX 2060 6GB, Ryzen 5700X, 32GB DDR4 RAM. Not the beefiest PC especially for AI rendering but I didn’t think it would take this long.

Again, is this like a first time setup deal, or do I have to download loads of GBs to generate one image?
 
  • Like
Reactions: Frindis and Pifanjr
So I followed a guide to get Flux working in Pinokio. Took a good 20 minutes to install everything. Finally, Gradio shows up and lets me input a prompt. I decided to do something simple, so I inputted “two guinea pigs frolicking in a flowery meadow, cute pastel art style”. Pinokio then proceeded to attempt to download about 30gb of files, and after 10 minutes the image was still not generated. I terminated the request.

Is that normal for each prompt, or is it like a one time thing? I seen in the command prompt something about developer mode needing to be turned on so Hugging Face can more efficiently store cache, so I did that and restarted my PC. I then tried to make an image again and it still happened.

I have a RTX 2060 6GB, Ryzen 5700X, 32GB DDR4 RAM. Not the beefiest PC especially for AI rendering but I didn’t think it would take this long.

Again, is this like a first time setup deal, or do I have to download loads of GBs to generate one image?

It has to actually download the Flux model first, which is a one-time download.
 
It has to actually download the Flux model first, which is a one-time download.
Got it, that’s kind of what I was thinking. It was confusing because when first installing it, I thought what I was doing was installing the actual model along with all the different software needed to make it run. Perhaps it was just the webui not the actual model. I suppose when I actually generated the image the first time, that’s when it downloaded the full model. I’ll try it again tonight and be more patient with it.
 
  • Like
Reactions: Frindis and Pifanjr

ZedClampet

Community Contributor
Got it, that’s kind of what I was thinking. It was confusing because when first installing it, I thought what I was doing was installing the actual model along with all the different software needed to make it run. Perhaps it was just the webui not the actual model. I suppose when I actually generated the image the first time, that’s when it downloaded the full model. I’ll try it again tonight and be more patient with it.
yeah, the installation and startup requires a lot of patience. Once everything is fully ready to go, you should be fine.

What a lot of people with low-end systems like us do is have it make small pictures and then let the AI make them bigger. There are different ways to do this. I'm still a noob. But in Forge if you go to one of the tabs, don't remember what it's called off-hand, there is an AI powered image enlarger.
 

ZedClampet

Community Contributor
I call this "Right Before Dinner"

full



The gigantic lion in a cage with part of it missing is a little terrifying. AI has a very consistent history of making what would be a great picture if only....

I was trying to have people visiting a Victorian menagerie, but I didn't like any of the pictures, so I thought I'd just post the terrifying lion right before he has dinner.

I've got half-a-dozen pictures similar to this. Flux doesn't want to make the lion fully encaged.
 

ZedClampet

Community Contributor
So what can you use to make good pic that you don't have to pay for? I've actually had an idea to make a visual novel using nothing but AI art. Ya know for shits and giggles. Free is so damned limited though.
Flux isn't limited at all. It sits on your PC and you can do whatever you want. Flux is free and the best there is, even better than the paid ones. If you use pinokio (see below), there is a separate program (all free and unlimited) you can install for visual novels that allows you to keep all your characters and environments consistent from one picture to the next. However, you can also just use flux for that. You just have to train it on the characters and whatever.

I don't know if you've read my more detailed response to your "why pinokio?" question, but pinokio is a one click install. Without it, you need to install a lot of different things that require some serious knowledge, like setting up Nvidia's CUDA library and various things that require knowledge of python and github.

Most AI is actually free and unlimited if you know where to look. Pinokio is a great place for all sorts of free AI
 
Last edited:
Keep getting a low VRAM error with Flux. In Pinokio I found something called “Forge”, which claims to work with Flux on as low as 3GB VRAM for 512x512 pics. I don’t care too much about size and resolution, I just want to finally make something. Installing it now, let’s see how it goes.
 
  • Like
Reactions: Pifanjr and Frindis
I call this "Right Before Dinner"
I'm assuming you don't mean they're going to eat the lion...

full



The gigantic lion in a cage with part of it missing is a little terrifying. AI has a very consistent history of making what would be a great picture if only....

I was trying to have people visiting a Victorian menagerie, but I didn't like any of the pictures, so I thought I'd just post the terrifying lion right before he has dinner.

I've got half-a-dozen pictures similar to this. Flux doesn't want to make the lion fully encaged.
That's Azlan, the cage is ceremonial.

I am more interested in what kind of vehicle that huge wheel belongs to.
 
By the way, Forge is what I told you to download. Sorry for the confusion.
You’re fine haha. I used a guide to install the regular fluxwebui and kept getting all of those errors. In Pinokio I just searched “vram”, and Forge was the first result. Definitely going to give it a better run when I have time, it was a lot of fun especially knowing it’s generated locally and not at some supermassive data center.
 
there is a separate program (all free and unlimited) you can install for visual novels that allows you to keep all your characters and environments consistent from one picture to the next
Consistency isn't my goal. I was thinking a parody VN where crappy AI art is the central plot and the AI characters know they're crappy AI art.
 
Last edited:
  • Like
Reactions: Pifanjr

ZedClampet

Community Contributor
j2AiFJn.png


After several tries with different topics this is the first one that was actually close to what I wanted without any oddball or grotesque anomalies using Flux.
It can be very frustrating. I was trying to make 10 pictures of a museum exhibit during a fund raising event (don't ask why), and there was always something wrong with the pictures. A picture would be 98 percent perfect, and then you'd notice someone's hand that looked like it had spent time in a garbage disposal. As much of an improvement as Flux is over previous generations, it still has a ways to go. At least most of the absolutely horrifying nightmare fuel has gone away.
 

ZedClampet

Community Contributor
Consistency isn't my goal. I was thinking a parody VN where crappy AI art is the central plot and the AI characters know they're crappy AI art.
Then Flux is your huckleberry. Actually any current gen AI would be, but you might as well use the free one that's on your PC.

If you want it to be really terrible, though, you should switch to Stable Diffusion 1.5.

Also, to you and @neogunhero, Forge seems to become corrupted the more you use it. If you are getting really bad pictures that aren't even what you are asking for, usually a reboot will do the trick.

This goes for using Distilled CFG as well. If you turn that up a little and all your images become blurry, it's time for a reboot. Sometimes regular CFG will work and sometimes it won't. No idea what is going on there. Something with Forge. One of these days I'm going to learn ComfyUI.