Ironhide's AI art gallery - feel free to exhibit your own art

How does it work? Do you literally just type something in and it generates a picture?

Albear wrote:How does it work? Do you literally just type something in and it generates a picture?

That's exactly how it works. With DALL-E I would type something like 'cubist oil painting of a fat man on a stool surrounded by Volkswagen memorabilia'.

This could be a great tool for bringing your dreams to life. My brother has a recurring dream about Danny Glover digging up his garden in a mini JCB. I might need to explore this!

hamm sandwich wrote:This could be a great tool for bringing your dreams to life. My brother has a recurring dream about Danny Glover digging up his garden in a mini JCB. I might need to explore this!

Isn't he a bit old for that gooseberry fool?

Albear wrote:How does it work? Do you literally just type something in and it generates a picture?

Depends on which AI art generator you use, some like Dall-e need you to specify things like composition and art style to get better results whereas the one I'm using (artflow.ai) lets you choose from several art styles and aspect ratios.

Midjourney is the most technically impressive one around but its very fiddly as you have to use Discord and type in console commands to set it going (plus there's a 25 image limit for free accounts, you have to pay $8 per month for more - I am tempted to try this though as I really enjoy messing around with ai art).

I asked Bing to draw an anime magical girl having a rap battle with Mr Blobby

Ironhide wrote:
Albear wrote:How does it work? Do you literally just type something in and it generates a picture?

Depends on which AI art generator you use, some like Dall-e need you to specify things like composition and art style to get better results whereas the one I'm using (artflow.ai) lets you choose from several art styles and aspect ratios.

Midjourney is the most technically impressive one around but its very fiddly as you have to use Discord and type in console commands to set it going (plus there's a 25 image limit for free accounts, you have to pay $8 per month for more - I am tempted to try this though as I really enjoy messing around with ai art).

There's also Stable Diffusion which you can run for free on your own PC, but it requires a good GPU and takes a bit of time to get working and then it just might not work for some reason and you need to troubleshoot that.
That's what made this nightmare for me:

I was actually trying to go much more extreme, but it's clear that whatever's not part of the model is very difficult to get it to make (or it just takes more effort or knowhow). So the knife is not plunged into her eye or her neck (and I definitely would not have shared that, or kept it on my computer for long! but you gotta experiment and find where the boundaries lie (well, some of them at least)), and the sky behind her is not filled with a dozen screaming eyes etc.

These cats look nowhere near as evil and murderous as I want them to:

YouTube

OldSoulCyborg wrote:
Ironhide wrote:
Albear wrote:How does it work? Do you literally just type something in and it generates a picture?

Depends on which AI art generator you use, some like Dall-e need you to specify things like composition and art style to get better results whereas the one I'm using (artflow.ai) lets you choose from several art styles and aspect ratios.

Midjourney is the most technically impressive one around but its very fiddly as you have to use Discord and type in console commands to set it going (plus there's a 25 image limit for free accounts, you have to pay $8 per month for more - I am tempted to try this though as I really enjoy messing around with ai art).

There's also Stable Diffusion which you can run for free on your own PC, but it requires a good GPU and takes a bit of time to get working and then it just might not work for some reason and you need to troubleshoot that.
That's what made this nightmare for me:

I was actually trying to go much more extreme, but it's clear that whatever's not part of the model is very difficult to get it to make (or it just takes more effort or knowhow). So the knife is not plunged into her eye or her neck (and I definitely would not have shared that, or kept it on my computer for long! but you gotta experiment and find where the boundaries lie (well, some of them at least)), and the sky behind her is not filled with a dozen screaming eyes etc.

These cats look nowhere near as evil and murderous as I want them to:

Aye that's what I use.

I've been trying to work out how to use that for a couple of hours, I can't get it to use my GPU as its an AMD one.

Got it to generate an image but it took ages and the result wasn't anywhere near what artflow generates in a few seconds.

Going to try importing an image from artflow and upscaling it with Stable Diffusion.

Made this with Stable Diffusion (Easy Diffusion - it's a bit odd, it runs in browser but takes a while to 'build' itself beforehand).

Mage holding crystal ball, drawn in the style of Hayao Miyazaki

Not bad considering I have no idea what I'm doing with it yet.

Been reading a few tips for getting better resuls from prompts, seems using a string of keywords is better in a lot of cases.

Stool bloke :lol:

twitter.com/WeirdAiGens/status/1647700412248498176

If you're interested in generating anime style images exclusively, Ironhide, you probably want to install one of the anime focused derivatives of Stable Diffusion rather than using the base model. (You can get very interesting results out of the base model but it's a jack of all trades.)

Stugene wrote:

Bumping this in case anyone has forgotten it.

site23 wrote:If you're interested in generating anime style images exclusively, Ironhide, you probably want to install one of the anime focused derivatives of Stable Diffusion rather than using the base model. (You can get very interesting results out of the base model but it's a jack of all trades.)

I was thinking about doing that as compared to what artflow can produce I'd say the results from SD a fairly poor in comparison.

I want to try doing some sci-fi themed stuff as well as the anime stuff, mainly to add some variety to.

I'm so glad I saved this thread to look through on a Monday :lol:

it's made my week already!

Agreed, it's one for the ages.

Getting the hang of it now after looking into how to use prompts more efficienctly.

kazanova_Frankenstein wrote:Agreed, it's one for the ages.

As is your avatar. :lol:

Ironhide's AI art gallery - feel free to exhibit your own art

Who is online

Affiliates