jaftt.bsky.social
@jaftt.bsky.social
Genesis
genesis-embodied-ai.github.io
December 21, 2024 at 3:09 AM
Are you illiterate?
December 21, 2024 at 3:07 AM
i got great news dude x.com/zhou_xian_/s...
x.com
x.com
December 20, 2024 at 10:54 AM
wouldnt that apply to comics and anime?
December 20, 2024 at 10:54 AM
the point is that its far more complex than prompting
December 19, 2024 at 8:54 PM
most artists dont. its not a coincidence that anime and comic books all have very recognizable art styles. is that theft?
December 19, 2024 at 8:53 PM
itll get better. also, the average length of a shot is like 3 seconds stephenfollows.com/p/many-shots...
How many shots are in the average movie?
Last week, I experienced the latest Transformers movie.
stephenfollows.com
December 19, 2024 at 8:52 PM
ai are can also be complicated. there are different tools like comfyui, ipadapter, controlnet, ic-light, etc. that help control the output. it can be VERY complex, like this one: openart.ai/workflows/da...
Manga Cosplay | ComfyUI Workflow | OpenArt
Created by: Datou: walkthrough video: https://www.youtube.com/watch?v=TSiNDm8q3vI&t=858s This workflow is only suitable for single-person scenes as it only replaces one face. If you want to use multi-...
openart.ai
December 19, 2024 at 12:34 PM
the point is that it can be consistent
December 19, 2024 at 4:44 AM
taking a photo is just painting with extra steps
December 19, 2024 at 4:43 AM
As opposed to drawing it from scratch for every single frame
December 17, 2024 at 11:16 AM
You literally do the same thing when you associate the characteristics of an apple with the word apple. Thats the whole point of this painting publicdelivery.org/wp-content/u...
publicdelivery.org
December 16, 2024 at 1:01 PM
Lmao. Shows how much you know about character design
December 16, 2024 at 12:58 PM
Compare the amount of time spent drawing it from scratch vs fixing a few buttons.
December 16, 2024 at 12:58 PM
If it works for art, it works for video

Proof: civitai.com/models/10346...

And flaws can be fixed with in painting or photoshop
Hunyuan Video - Arnold Schwarzenegger - LoRA - v1.0 | Hunyuan Video LoRA | Civitai
First attempt at a HunyuanVideo LoRA: Arnold Schwarzenegger - Trigger is "ohwx person" - include "ohwx person" in your prompt to call up the Arnold...
civitai.com
December 16, 2024 at 12:57 PM
Looks like the same character in all 4 images
December 12, 2024 at 12:06 AM
The point is that the character is consistent. Minor details can be fixed with inpainting and img2img
December 12, 2024 at 12:05 AM
Yes. If you want to understand it, read the papers
December 12, 2024 at 12:05 AM
You do that too. How do you know what an apple is but not تفاح? Because you associate apples with the word "apple" but have never seen the word تفاح before.
December 12, 2024 at 12:04 AM
Finding the average does not allow for zero shot learning, which LLMs can do cs229.stanford.edu/lectures-spr...
cs229.stanford.edu
December 10, 2024 at 10:34 PM
If it was a weighted average, it couldn't perform zero shot learning. But it does: cs229.stanford.edu/lectures-spr...
cs229.stanford.edu
December 10, 2024 at 10:24 PM
How
December 10, 2024 at 10:21 PM
AI helps the human soul

First legally recognized nonbinary person with disabilities writes book with ChatGPT: www.wired.com/story/the-us...
How One Author Pushed the Limits of AI Copyright
Elisa Shupe was initially rebuffed when she tried to copyright a book she wrote with help from ChatGPT. Now the US Copyright Office has changed course—but there’s a catch.
www.wired.com
December 10, 2024 at 10:21 PM