For Those Experimenting With AI

[LHerrmann]

Yes, this one is generating everything without limitations.
If you buy it, I recomment using the models "Flux", "Flux reality" and "Shuttle difusion"

 

[stockingfeetrbest]

Yes, this one is generating everything without limitations.
If you buy it, I recomment using the models "Flux", "Flux reality" and "Shuttle difusion"

I'm confused why this isn't working any better than some other programs I tried. So far the only one with images that aren't completely wrong or distorted is Chat GPT. In the first pic I asked it to show a man restrained on a padded table with his ankles and wrists in stocks at each end of the table, he is wearing a nylon bodystocking and 2 women in black hooded robes are tickling his feet and armpits. The second pic I asked it to show a jester tickling a Queen's stocking feet.

Is there a way to edit these after they are created? With ChatGPT you can tell it to change things. I don't see that option here. These 2 were with Flux.

Update: I have tried the other 2 models you suggested with no better results. I have cancelled my subscription for now even though I am out $19. I may ask for a refund but am not holding my breath. What are you doing different to get the results you posted? It isn't giving me anything close to what I am describing. The last one I tried was a spy with her nylon feet in stocks. It showed me a pic of a woman sitting on the floor and she had 3 legs. LOL

 

[LHerrmann]

I'm confused why this isn't working any better than some other programs I tried. So far the only one with images that aren't completely wrong or distorted is Chat GPT. In the first pic I asked it to show a man restrained on a padded table with his ankles and wrists in stocks at each end of the table, he is wearing a nylon bodystocking and 2 women in black hooded robes are tickling his feet and armpits. The second pic I asked it to show a jester tickling a Queen's stocking feet.

Is there a way to edit these after they are created? With ChatGPT you can tell it to change things. I don't see that option here. These 2 were with Flux.

Update: I have tried the other 2 models you suggested with no better results. I have cancelled my subscription for now even though I am out $19. I may ask for a refund but am not holding my breath. What are you doing different to get the results you posted? It isn't giving me anything close to what I am describing. The last one I tried was a spy with her nylon feet in stocks. It showed me a pic of a woman sitting on the floor and she had 3 legs. LOL

Well, there are some things you can try:

• First of all, try to learn some basic prompt-Engineering skills. Which means, you should write your prompt so the AI can understand what you are meaning. For example: If you write something like "tickling" some models don't even know what this is, because it is nothing you can "show directly", so you have to describe it more in detail. When you say, "the women tickle his feet" you should add "the women tickle his feet by touching the soles of the feet..." little difference in the prompt, but high impact. The more concrete you are, the better result you get. For example: AI doesn't know what tickling is: So you have even to describe further: Touching the body, tickling the body... the man is closing his eyes, laughing hysterically..."
• 2nd use "advanced settings" the most models have advanced settings: You can play with steps and prompt guidance. By using Flux, Shuttle diffusion; you can increase the image size, which automatically increases the steps and qualitycontrol of your output.

 

[stockingfeetrbest]

Well, there are some things you can try:

• First of all, try to learn some basic prompt-Engineering skills. Which means, you should write your prompt so the AI can understand what you are meaning. For example: If you write something like "tickling" some models don't even know what this is, because it is nothing you can "show directly", so you have to describe it more in detail. When you say, "the women tickle his feet" you should add "the women tickle his feet by touching the soles of the feet..." little difference in the prompt, but high impact. The more concrete you are, the better result you get. For example: AI doesn't know what tickling is: So you have even to describe further: Touching the body, tickling the body... the man is closing his eyes, laughing hysterically..."
• 2nd use "advanced settings" the most models have advanced settings: You can play with steps and prompt guidance. By using Flux, Shuttle diffusion; you can increase the image size, which automatically increases the steps and qualitycontrol of your output.

What about the issue with the Queen's feet looking all distorted. This happens with every AI platform except ChatGPT. Sometimes they literally look like cucumbers.

Here is another botched attempt. Here is exactly what I wrote using Flux.

"a female spy laying flat on a table with her feet locked in restraints at the end of the table. two women are tickling her feet by touching the bottoms of her feet with their fingers. she is laughing hysterically. she has pink socks on her feet."

There is no way the AI can be that bad!

 

[stockingfeetrbest]

I give up on AI image generation for now. I have wasted money on multiple AI programs and none of them work except for ChatGPT and they suck. Way too restrictive.

 

[theatergoer]

What about the issue with the Queen's feet looking all distorted. This happens with every AI platform except ChatGPT. Sometimes they literally look like cucumbers.

Here is another botched attempt. Here is exactly what I wrote using Flux.

"a female spy laying flat on a table with her feet locked in restraints at the end of the table. two women are tickling her feet by touching the bottoms of her feet with their fingers. she is laughing hysterically. she has pink socks on her feet."

There is no way the AI can be that bad!

It's how tickling scenes look in Chernobyl. 😉

 

[stockingfeetrbest]

ChatGPT has lost it's mind. I ask for the simplest things and it violates policy but today I asked it to "create an image of a woman in a fun house tickling feet in a wall". Then I simply added "the hole in the wall is padded" and got the 2nd pic. NO other prompts which completely contradicts what others have said about the AI not knowing what tickling is without substantial descriptions. Pretty impressed with the woman's appearance, outfit and the look on her face. Lucky lady is having some serious fun. For now I am imagining the feet are in black nylon toe socks.

I'm starting to think the filters get bypassed sometimes at random. I asked it to "add orange nylons to the feet" and got the policy violation. I asked it "the feet have socks on" still a no. But it's ok for her to have Wolverine hands tickling a pair of feet insane? WTF?

 

[Yanks129]

ChatGPT has lost it's mind. I ask for the simplest things and it violates policy but today I asked it to "create an image of a woman in a fun house tickling feet in a wall". Then I simply added "the hole in the wall is padded" and got the 2nd pic. NO other prompts which completely contradicts what others have said about the AI not knowing what tickling is without substantial descriptions. Pretty impressed with the woman's appearance, outfit and the look on her face. Lucky lady is having some serious fun. For now I am imagining the feet are in black nylon toe socks.

I'm starting to think the filters get bypassed sometimes at random. I asked it to "add orange nylons to the feet" and got the policy violation. I asked it "the feet have socks on" still a no. But it's ok for her to have Wolverine hands tickling a pair of feet insane? WTF?

Amazing artwork

 

[Ticklefeather00]

If you're going to try and do more complex images with AI, be prepared to kiss a lot of frogs and by that I mean a large number of renders. Also be prepared to spend a little because the free ones I've tried are awful. Have you tried ideogram.ai?

 

[Kalamos]

If you're going to try and do more complex images with AI, be prepared to kiss a lot of frogs and by that I mean a large number of renders. Also be prepared to spend a little because the free ones I've tried are awful. Have you tried ideogram.ai?

I spent some money to bypass waits, but in the end it wasn't really worth it - at least for the generators I used.
I tend to avoid generating contents from scratch though, and I usually create middle frames from existing images.