Page 2 of 2
Re: Having fun with the AI
Posted: Fri Oct 11, 2024 5:12 pm
by superheroinelinks
Re: Having fun with the AI
Posted: Fri Oct 11, 2024 7:58 pm
by knottyguy
Bladebur wrote: ↑1 year ago
Batgirl Walls.png
I don't usually like Batgirl costumes without the leggings but this one just works somehow!
Re: Having fun with the AI
Posted: Fri Oct 11, 2024 10:11 pm
by Malice
knottyguy wrote: ↑1 year ago
Bladebur wrote: ↑1 year ago
Batgirl Walls.png
I don't usually like Batgirl costumes without the leggings but this one just works somehow!
Sorry!

Re: Having fun with the AI
Posted: Sat Oct 12, 2024 8:17 am
by Bladebur
knottyguy wrote: ↑1 year ago
I don't usually like Batgirl costumes without the leggings but this one just works somehow!
My preference is the opposite, but I admit it's hard to get them right. There are lots of nuances in the art of showing some leg!
Re: Having fun with the AI
Posted: Sat Oct 12, 2024 8:21 am
by Bladebur
I wanted Batgirl and Catwoman in a double death trap, and here they are!

- Batgirl Catwoman Cylinder.png (1.93 MiB) Viewed 3256 times
Not 100% Batman-1966-ish, but close enough. I wanted question marks in that suit but the AI goes wonkers with that and I'm not going to paint them by hand XD
Re: Having fun with the AI
Posted: Sun Oct 13, 2024 6:52 am
by Bladebur
I've also been having fun with voice AI. There is this model at
https://swivid.github.io/F5-TTS which can take a clip of some voice and then produce another one with a different text. It works pretty well! Here is a fun example with Eartha Kitt's Catwoman.
(I wish this forum could embed audio just like you can do with images, by the way)
https://thunder-girl.org/catwoman-demo.wav
Re: Having fun with the AI
Posted: Sun Oct 13, 2024 7:28 am
by Malice
Bladebur wrote: ↑1 year ago
I've also been having fun with voice AI. There is this model at
https://swivid.github.io/F5-TTS which can take a clip of some voice and then produce another one with a different text. It works pretty well! Here is a fun example with Eartha Kitt's Catwoman.
(I wish this forum could embed audio just like you can do with images, by the way)
https://thunder-girl.org/catwoman-demo.wav
Oooooh yeaaaaaaaaaaaaah.

Re: Having fun with the AI
Posted: Sun Oct 13, 2024 7:47 am
by Bladebur
Malice wrote: ↑1 year ago
Bladebur wrote: ↑1 year ago
I've also been having fun with voice AI. There is this model at
https://swivid.github.io/F5-TTS which can take a clip of some voice and then produce another one with a different text. It works pretty well! Here is a fun example with Eartha Kitt's Catwoman.
(I wish this forum could embed audio just like you can do with images, by the way)
https://thunder-girl.org/catwoman-demo.wav
Oooooh yeaaaaaaaaaaaaah.
Also, the model is completely uncensored, so no problem with being dirty (you can see the model has a bit more problems since that's probably going below its training, but it's still somewhat believable and you'll probably get something better with many attempts and retries).
https://thunder-girl.org/catwoman-spicy.wav