Photorealistic AI Generated Images

#fluff

Working on a new character lora:

image (5).pngimage (3).png
 

Attachments

  • image (2).png
    image (2).png
    1.3 MB · Views: 8
  • image.png
    image.png
    1.3 MB · Views: 9
  • Love
Reactions: jcjc and Casshern2
I made some maids. :D Then I tried a very simple prompt to let the LTX model imagine the rest. It took the boring approach, but at least there is movement. As I've seen when you leave a lot for the model to do, lots of zooms. I'll try some more detailed prompts later to see what it does. I did generate 2 videos per image but only kept the better looking ones.

00000-154014950.jpg00000-1256921174.jpg00001-2359775246.jpg00002-3490512563.jpg00005-1422046878.jpg00006-2206566337.jpg00008-2604518352.jpg00009-2604518355.jpg

Switching to FileJoker for two reasons:
  • keeps me compliant with posting files
  • Akiba doesn't support embedding videos, so the Mega embed mark doesn't work anyway, have to use the mp4 path
  • because I'm using he mp4 path, browsers will play at full browser size which stretches the true 1024x576 video frame
Simple prompts for maids: A pretty Japanese maid looks happy and smiles at me.
https://filejoker.net/vezbhnzrvanw/maids.mp4 (21.8 Mb)

 
Last edited:
  • Love
Reactions: jcjc