Artificial Intelligence
- Started
- Last post
- 1,637 Responses
- cherub0
Well... have we?
- neverscared4
Albert Einstein in Stranger Things
#StableDiffusion #AIart
- kalkal2
MJ to start with, DALLE inpainting for the clothing changes, CodeFormer for the face, and a ERSGAN upscale
- https://i.imgur.com/… << straight out of MJkalkal
- Wow, flawless. Finally a workable workflow!
Scary good? More like very challenginggrafician - In a way this challenges again the idea of "designers should code?" guess the answer is very affirmative these daysgrafician
- doesn't require any codingkalkal
- well.. not on my part lolkalkal
- Sure, not in these cases, but if one needs more control over the process...I mean what Disco Diffusion was like python code?
That sort of control tweakinggrafician - You do edit python to control peramterrs of disco diffusion but it's not really programming per say, more very poor uxkalkal
- grafician0
"Running your own A.I. Image Generator with Latent-Diffusion
Run your own text to image prompts with CUDA, a bunch of disk space, and an insane amount of memory."
"In this post, we’ll look at getting setup with running your own A.I. image generator. You need a Linux system with a CUDA enabled card to get this working through to the end. I did get the basic txt-2img script working in Windows. Unfortunately, for the July released retrieval-augmented diffusion models, you need scann in order to index the openimages dataset, which is only available on Linux. WSL worked fine for me."
"Here’s roughly what I gather you’ll need:
Linux (WSL fine)
11.3 CUDA graphics
>= 32GB RAM
50GB disk space (~30GB without openimages)
3xxx series card (I am just guessing here. This is all being done with a 3090 FE)"
- grafician0
Comparison between Midjourney and Difussion:
prompt: "angel girl face portrait in the style of iris van herpen"
- grafician0
- Discussions about copyrights starting to pop-up...grafician
- Meanwhile, I have nothing to add to the discussion, just enjoying it for what it is. If it goes away, too bad but I'm enjoying it for now.kalkal
- In these discussions I see many people don't actually know how it works, in the first placegrafician
- Yeah, that's true, they think it's a smart automatic photobash or something.kalkal
- Someone told me that on MJ, they saw someone in the 10k club who was complaining about MJ not removing watermarks as though it's a image search or somethingkalkal
- Ultimately, I think many probably don't want to know how it works beyond "that looks like my artwork, my artwork was in the pool of data, they stole my work"kalkal
- I've seen some hypocriticism too, desginers saying they want the option to pull their work from the model and justify using MJ by saying they only use promptskalkal
- from long dead artists (an arbitrary line that they are creating for themselves) and ignoring the fact that the model pulls from all data it holdskalkal
- And then cherry on the top, using image prompts that look like they grabbed them off of instagramkalkal
- In conclusion... I guess I do have something to say, I don't care enough to argue with someone about it though lolkalkal
- Yup, samegrafician
- neverscared4
works fine , where is my selfdriving car already...
- rzu-rzu1
- this is the original twitter post with more details:
https://twitter.com/…grafician
- this is the original twitter post with more details:
- neverscared18
- They’ll just get you to describe what they want and pay you half.Chimp
- There's a new job emerging for copywriters who are good at this kind of thing, it's a similar skill to SEO writing. It's titled 'Prompt Engineer'.microkorg
- Oh snapYakuZoku
- Many "prompters" already on instagram with "Open for comissions" in their bio...grafician
- Will not replace, but will steal some sizable share of workdrgs
- i doubt it will steal a lot .. it will be just an add on tool for every designer who needs and wants it in their repertoire to find solutions to design problemsneverscared
- feeding your own 3d rendering, illu, animation fru the synthetic Intelligence will be nice .... the other stuff is just unoriginal poser stuff...and will be atneverscared
- the margins where it belongs to puny skill space...neverscared
- margins where it belongs to puny skill space.neverscared
- oopsneverscared
- PhanLo0
LoL
-
- grafician-2
- "A subreddit dedicated to Art and projects created using Visions of Chaos. Get Started: https://softology.pr…grafician
- _me_2
- this video is a shorter version (or cut) from the other one posted by rzu-rzu a few posts before
https://qbn.com/repl…grafician - more like cut and reposted (stolen) but okaygrafician
- this is the original one: https://twitter.com/…grafician
- thanks cunt._me_
- you're welcome bitchgrafician
- this video is a shorter version (or cut) from the other one posted by rzu-rzu a few posts before
- kalkal1
Every useful ML manipulation I'm aware of used here -
Raw Stable Diffusion output -
- fingers..shapesalad
- saladkalkal
- btw, https://i.imgur.com/… 12 mins to generate 48 512x704 images locally on my 3080ti, same resolution as this image started as <<<<kalkal
- and an outdated 8gb unoptimised model at thatkalkal
- microkorg-1
Woman answers questions from the grave thanks to AI.
- kalkal0
The Stable Diffusion model leaked and can be run on your own rig locally early
- What nownb
- Freaky NSFW shit?kalkal
- but it's going to be public and free soon anyway: https://stability.ai…uan
- could be an unstable version or some crypto miner scam :)uan
- If it is, does a pretty good job of working fine, aside from the fact it's a build from June and not quite as goodkalkal
- wish I had the time and hardware to play it :/uan
