Ah, i thought the lyrics were AI generated at first and i was very impressed
Just started messing with Dance Diffusion for audio, really interesting.
I put in a piece of music i made and it added (nonsensical) lyrics which is pretty cool, it also changed the music somewhat, it had electronic drums it changed them to acoustic drums.
Gonna put in some of my sample libraries to make it come up with completely new sounds.
well development certainly hasn't slowed down.. mashed together some models and it turned out pretty great.. no photoshop, had to fix an eye via inpainting that's all.
@vurt
What model and/or prompts are you using? While I can get amazing looking people. Anything I try to do more than a torso and up shot turns into a nightmare of random extra limbs.
-We don't control what happens to us in life, but we control how we respond to what happens in life.
-Hard times create strong men, strong men create good times, good times create weak men, and weak men create hard times. -G. Michael Hopf
Disclaimer: Post made by me are of my own creation. A delusional mind relayed in text form.
@vurt
What model and/or prompts are you using? While I can get amazing looking people. Anything I try to do more than a torso and up shot turns into a nightmare of random extra limbs.
What are you using?
DALL-E is a bit crap compared to Midjourney, IMOOOO!
i use whatever models i find cool on https://civitai.com/ i merged 3, then i merged 3 with those.
protogenX53Photorealism_10 etc
i use stable diffusion webui.
extra limbs yes everyone gets those, it's important to configure your negative prompts to get rid of it, also scale is important, if its too high or too low you will get more of it.
in my negative prompt for these i had:
cartoon, anime, cartoonish, 3d render, extra limb, lowres, bad anatomy
so it's pretty short. i add more if i see problems, some people have a whole page of negative prompts, i don't think that works that great, it makes the images boring and lacking. i should have added "clone" or "duplicate" though because i do have one that sticks out and is ugly here.
all A.I; images, voice, music, writing (by ChatGPT).
just learned how to steal anyone's voice and render sentences with it quality is superb too, much better than super expensive commercial text-speech, i just listened to a comparison and this one sound way more natural.
and this was a sloppy model i made, only 1 file of 14 seconds of speech instead of the suggested 4-5 files. what i did was to speed up the voice but not the pitch in Soundforge, so a lot of data in one file, if i make 4 more files using that technique that voice is going to be almost without faults whatsoever (i think, because it's already good). the narrator voice is from George Guidall.
That is super awesome, going to have to play with it!
the text to voice would be awesome for our tabletop game for 'voiceover/description' stuff I could stick in here and there. Just that collab for the voice thing is horribly laid out (no collapsable cells, output and code is in the same cell), going to try a local install of it.
-We don't control what happens to us in life, but we control how we respond to what happens in life.
-Hard times create strong men, strong men create good times, good times create weak men, and weak men create hard times. -G. Michael Hopf
Disclaimer: Post made by me are of my own creation. A delusional mind relayed in text form.
@vurt
Thanks vurt...
Everytime I stop obsessing over some AI gen stuff with all my free time. You pop up and point out models or datasets I hadn't noticed and spend way too much time generating random crap
Right now I am using Automatic111's SD WebUI launcher since I got the 4080. Are you using a different SD webUI for the images (and/or other stuff)? If so any others you recommend over it?
-We don't control what happens to us in life, but we control how we respond to what happens in life.
-Hard times create strong men, strong men create good times, good times create weak men, and weak men create hard times. -G. Michael Hopf
Disclaimer: Post made by me are of my own creation. A delusional mind relayed in text form.
I also use the webui, with various models i find on the net. but i will switch to invokeAI since it's just much better for making immensely large images via inpainting. i just haven't gotten around to it yet. i'm also gonna try to learn to better do my own models.. right now i'm mostly just mixing models that i find, and i do that very randomly just to see what happens. like 80% of my time spent with this is just experimenting.
Poor ducky at least he found a bit chocolate to snack on and even though it tasted a bit off, he at least had a nice story to tell
Added the above to ChatGPT (you can keep any particular session which ChatGPT will remember and continue from).
Quote:
Input: Poor ducky. at least he found a bit chocolate to snack on and even though it tasted a bit off, he at least had a nice story to tell.
ChatGPT: Yes, poor PumpAction. But he didn't let his misfortune to the toilet bowl get him down for too long. He decided to make the best of the situation and began to look for snacks. To his surprise, he found a piece of chocolate floating in the water.
Excitedly, he paddled over to it and took a bite. But to his dismay, the chocolate tasted a bit off. It had a strange, chemical-like flavor. Nevertheless, he finished the chocolate and decided it was still a nice treat.
From that day on, PumpAction had a new hobby: searching for strange snacks in his toilet bowl home. And even though it wasn't the pond of his dreams, he was content knowing that he had a unique and funny story to tell all his duck friends.
I turned 7 character images to animated 2.5D
it's kind of like experiencing magic, one second its a standard 2D image, next it's something much different and life-like.. i can hardly believe it can gather and add so much to just ordinary 2D.
You generate that audio?
If so did you train your own or use a downloaded ckpt?
-We don't control what happens to us in life, but we control how we respond to what happens in life.
-Hard times create strong men, strong men create good times, good times create weak men, and weak men create hard times. -G. Michael Hopf
Disclaimer: Post made by me are of my own creation. A delusional mind relayed in text form.
messing with image-to-image right now. i rendered a studio Ghibli anime-style about "ice cream tornados and monsters" over a random youtube music video. Looks cool and it can be hilarious, but there's no consistency, looks super fluid and funky though, you can make impressive music videos with it.
Spoiler:
its a part from this video:
if a model is trained on material that isn't random but very specific(click) it could be done.. its worked on in all areas, 3D, music, voice, writing/science, image/video. It's the biggest thing happening for computers since the internet.
"Enlightenment is man's emergence from his self-imposed nonage. Nonage is the inability to use one's own understanding without another's guidance. This nonage is self-imposed if its cause lies not in lack of understanding but in indecision and lack of courage to use one's own mind without another's guidance. Dare to know! (Sapere aude.) "Have the courage to use your own understanding," is therefore the motto of the enlightenment."
All those impressionists on Fiverr becoming expendable in 3..2..1..
Looking forward to the modded versions of future Rick & Morty eps with Roiland's voiceover reinserted
Signature/Avatar nuking: none (can be changed in your profile)
You cannot post new topics in this forum You cannot reply to topics in this forum You cannot edit your posts in this forum You cannot delete your posts in this forum You cannot vote in polls in this forum
quality is superb too, much better than super expensive commercial text-speech, i just listened to a comparison and this one sound way more natural.
at least he found a bit chocolate to snack on and even though it tasted a bit off, he at least had a nice story to tell 
i can hardly believe it can gather and add so much to just ordinary 2D.
