We made the first comic fully generated by an AI using DreamBooth finetuning of Imagen at @GoogleAI. We use only 4 input images of our cute little character Anselmo (drawn by my friend @IsaArduz). (thread) 1/N #DreamBooth #AIArt #Imagen #stablediffusion #midjourney #AIArtwork

Sep 12, 2022 路 8:15 PM UTC

16
80
25
431
These are the input images for our character Anselmo. We generate a fully-fledged comic with Anselmo in new poses, with different accessories and even with text and speech bubbles automatically drawn by the diffusion model! (just prompt for "a [V] cartoon saying XYZ"!) 2/N
2
1
30
If this is not the first instance of a comic drawn by an AI please let me know here 馃榾 - have not been able to find any others. 3/N
7
18
If you are interested in our work, you can read more here dreambooth.github.io/ collaborators: @AbermanKfir @jampani_varun Yuanzhen Li @MikiRubinstein Yael Pritch) END
3
1
18
Replying to @natanielruizg @GoogleAI @IsaArduz
Wait, the whole storyline or individual frames?! Not sure if it qualifies but I was playing with it in the same direction! Didn't have enough memory for a decent textual inversion but known characters are fun as well
hakum @_ohakum
2 Sep 2022
Fun with stable diffusion - lonely marvel heroes enter Edward Hopper paintings pdf: shorturl.at/efoUZ #stablediffusion #Marvel #AIArtwork
1
5
Super cool!
1
Replying to @natanielruizg @GoogleAI @IsaArduz
i think it鈥檚 appropriate that the first three lines show gluttony, sloth, and pride
1
5
Very cool interpretation 馃槈
2
Replying to @natanielruizg @GoogleAI @IsaArduz
Were you restricted to only 4 training images, or could you have used even more e.g. so that the AI would have an idea of what he looked like from behind?
1
3
We could use more! And yes sure we could also have different poses
1
3
more replies
Replying to @natanielruizg @GoogleAI @IsaArduz
Is there anyway, where we can get access to DreamBooth?
1
1
There are unofficial implementations on Stable Diffusion that are available!
Load more