playing with AI image art (not for submission)

[joy_of_cooking]

Started playing with mage dot space thanks to @djrip in @MediocreAuthor 's visual inspirations thread.

The results are pretty funny. I asked for a woman holding a glass of wine and got a woman with a glass of wine in each hand, sitting in front of a table holding two more glasses of wine and a clutter of what looked like empty glassware.

Anyone having better luck? Tips for the newbie?

 

[MediocreAuthor]

Started playing with mage dot space thanks to @djrip in @MediocreAuthor 's visual inspirations thread.

The results are pretty funny. I asked for a woman holding a glass of wine and got a woman with a glass of wine in each hand, sitting in front of a table holding two more glasses of wine and a clutter of what looked like empty glassware.

Anyone having better luck? Tips for the newbie?

I've tinkered around with Ai art, and my best advice is "Manage your expectations" lol

I've had some success, but far more failure

 

[AlinaX]

Let's see if this works:
Apparently not.
mage dot space/c/88247d43d6424f5cbe2b473d5ae3d665
mage dot space/c/88eeed6a03fa4491a00179496a989cb8
Very unsafe...

 

[NoTalentHack]

Here’s some of what I’ve gleaned so far:

https://forum.literotica.com/threads/using-ai-to-create-“photos”-of-my-fictional-characters.1583836/post-97076256

If you go backwards and forwards in that thread done, you can find additional tips.

 

[alohadave]

This is one of the better AI interfaces that I've seen. The advanced options are pretty good on the free side, giving you a fair bit of control. And with unlimited generations, you can experiment to your heart's content.

It uses StableDiffusion (not surprisingly since it's fully open source), and you have options for which models to use.

The negative prompts will help refine your image by telling it what you don't want.

For the positive prompt, don't try to overload with a million different keywords, the more you have, the less importance each one has. You can put () around a prompt and that will increase it's strength.

Play around with SDXL, as that's the newest and tends to come up with better images with the least tweaking.

For $15/month for the paid option isn't too bad and is in line with other online AI service. And it gives you access to a lot of Loras and additional models to tweak with.

Overall, pretty good.

 

[onehitwanda]

One of my three in-the-works characters. Meet Cat, 27, perpetually single, and utterly enamored with her (petite, blonde) hairdresser.

https://image.lexica.art/full_jpg/de32741e-08a9-45d4-b01b-ebe91145d6b2

 

[MrPixel]

The results are pretty funny. I asked for a woman holding a glass of wine and got a woman with a glass of wine in each hand, sitting in front of a table holding two more glasses of wine and a clutter of what looked like empty glassware.

I used to know a couple of ladies like that. They thought it gauche to drink straight from the bottle.

 

[onehitwanda]

I used to know a couple of ladies like that. They thought it gauche to drink straight from the bottle.

https://media.tenor.com/E0A77KoErX8AAAAM/jessica-chastain.gif

still one of my favourite gifs ever.

 

[Fatdog25]

Started playing with mage dot space thanks to @djrip in @MediocreAuthor 's visual inspirations thread.

The results are pretty funny. I asked for a woman holding a glass of wine and got a woman with a glass of wine in each hand, sitting in front of a table holding two more glasses of wine and a clutter of what looked like empty glassware.

This sounds like the start of one of those "Yup, that's me..." stories.