Joined Jul 2022
Wopu
boosted
And the movie itself: https://youtu.be/V0ODG8bFme0
I'm not using Invidious links because they deserve your view.
For the love of Fuck, if you have 2 hours of spare time, there is an incredible full length movie done in Source Film Maker (SFM) called "Emesis Blue".
An animated Horror Thriller that manages to keep me on edge the whole duration of the plot, and relies on ZERO jumpscares to scare you.
Here is a "teaser" so you can get familiar with the theme: https://youtu.be/Qzqa527mVx8
Wopu
boosted
SHITTY STABLE DIFFUSION AI GUIDE FOR CPU ONLY POORS
>UPDATED 3/9/2023
Alright, I'm gonna spill my secrets on how I got this AI thing to work with only the CPU.
Before you go any further, I recommend you do this on a PC / server you're not actively using. If you're poor, and only have one computer you do everything with, you can follow the cputard guide here so you can at least get something that works:
https://rentry.org/cputard
Or you can dual boot and follow the rest of this guide on a separate linux partition. I only say all of this because I don't like shitting up my main desktop, and considering how memory intensive this all is, you wouldn't want a whole GUI running beside it too.
You're gonna need docker with docker-compose, a pretty good CPU (you have one of those right?) and lots of RAM. The memory thing is no joke, like this thing will hog up like 3-4 chrome tabs worth of RAM which means 8gb won't cut it. You'll want 16gb and above or else you'll be stuck in swap hell where you can't do jack shit.
Got docker running? Perfect, now clone this git repo:
git clone https://github.com/AbdBarho/stable-diffusion-webui-docker
once cloned, cd into the folder and follow the setup guide here
https://github.com/AbdBarho/stable-diffusion-webui-docker/wiki/Setup
You can find the links for the waifu generator models here:
https://rentry.org/voldy#-guide-
And plenty of models here:
https://huggingface.co/models?other=stable-diffusion
My personal favorites are anything-v4.0 and pastel-mix. Also any model on hugging face can be downloaded with curl or wget (just open the model repo, click the model, right click the "download" hyperlink and copy it for your terminal).
put the .ckpt file you want in data/StableDiffusion/
and bring up the AI and the UI with:
docker compose --profile auto-cpu up --build
It takes a few minutes but once its up you can connect to the webui over your LAN using
http://:7860
Enjoy your anime waifus. No I will not answer your question about the weird error you got (its probably your fault anyways). If I missed anything in this post TOUGH SHIT SORRY BUDDY (ok just reply to this and I'll edit the post with the correct info).
Huge thanks to @skylar @mkultra and @fluffy for the prompts and pointers when I started out. If I missed anyone else, thank you as well!
Alright, I'm gonna spill my secrets on how I got this AI thing to work with only the CPU.
Before you go any further, I recommend you do this on a PC / server you're not actively using. If you're poor, and only have one computer you do everything with, you can follow the cputard guide here so you can at least get something that works:
https://rentry.org/cputard
Or you can dual boot and follow the rest of this guide on a separate linux partition. I only say all of this because I don't like shitting up my main desktop, and considering how memory intensive this all is, you wouldn't want a whole GUI running beside it too.
You're gonna need docker with docker-compose, a pretty good CPU (you have one of those right?) and lots of RAM. The memory thing is no joke, like this thing will hog up like 3-4 chrome tabs worth of RAM which means 8gb won't cut it. You'll want 16gb and above or else you'll be stuck in swap hell where you can't do jack shit.
Got docker running? Perfect, now clone this git repo:
git clone https://github.com/AbdBarho/stable-diffusion-webui-docker
once cloned, cd into the folder and follow the setup guide here
https://github.com/AbdBarho/stable-diffusion-webui-docker/wiki/Setup
You can find the links for the waifu generator models here:
https://rentry.org/voldy#-guide-
And plenty of models here:
https://huggingface.co/models?other=stable-diffusion
My personal favorites are anything-v4.0 and pastel-mix. Also any model on hugging face can be downloaded with curl or wget (just open the model repo, click the model, right click the "download" hyperlink and copy it for your terminal).
put the .ckpt file you want in data/StableDiffusion/
and bring up the AI and the UI with:
docker compose --profile auto-cpu up --build
It takes a few minutes but once its up you can connect to the webui over your LAN using
http://:7860
Enjoy your anime waifus. No I will not answer your question about the weird error you got (its probably your fault anyways). If I missed anything in this post TOUGH SHIT SORRY BUDDY (ok just reply to this and I'll edit the post with the correct info).
Huge thanks to @skylar @mkultra and @fluffy for the prompts and pointers when I started out. If I missed anyone else, thank you as well!
@kaiservenom @Tfmonkey It's a shame Animation in the collective mind is relegated to "being for kids". Whatever Live Action can do, the Animation medium can also do it and make it much more expressive.
Wopu
boosted
Wopu
boosted
Wopu
boosted
@mutageno the shadowed hand behind the right chick is a bit uncanny but this AI is on the right path.
Man, it took me a whole hour to mentally do basic (pointer) arithmetic...
I couldn't for the life of me figure out the free space condition. A page later there is the definition of a function that gives you the length of a str variable then it hit me:
Storage allocation uses a Stack. allocp points to the first element of allocbuf, which is last object inserted. In effect, free space is ALLOCSIZE - (allocp - allocbuff) HENCE the fit condition.
To free pointer p, the condition is to be in allocp
"How I Programmed My Own AI Girlfriend": https://inv.riverside.rocks/watch?v=dKFnJCtcfMk
A video from a Youtuber describing his process of building an AI girlfriend, from the training samples to choosing the right Text-to-Speech/Deepfake voice for his needs, and passing through the choice of language models.
Video description links the Github projects he used.
A neat video on a topic of interest on the Saturday show.
Last note: The link directs to a Youtube video from a front-end
Wopu
boosted
@Tfmonkey Is it me or your thumbnail is having an anxiety attack?
Wopu
boosted
Wopu
boosted
Fans do a better job at making content than Toyotaro. Fan manga is just better. Dragon Ball Hakai and Dragon Ball Kakumei are other good ones to keep up with.
Joined Jul 2022
