Stable Diffusion How To Install And Run On ANY GPU! Text To Image Ai

preview_player
Показать описание
Anaconda

Stable Diffusion

Stable Diffusion Model

SD Helper Files + OptimizedSD

Commands
Рекомендации по теме
Комментарии
Автор

Dude this is the only tutorial for this that I've found so far that isn't a "draw the rest of the owl" type tutorial. Why the hell would an expert in running AI environments need a tutorial lol? Thanks for making a tut that actually explains the required prerequisites before just jumping in to a bunch of unexplained processes and programs I don't have yet!!

businessmonkey
Автор

If you have git issue:

1- conda install -c anaconda git

2- conda env remove -n environment.yaml

3- Delete the folder:

4-reinstall environment with:

conda env create -f environment.yaml

Thanks for the video

fallency
Автор

--prompt [prompt] - (This is the command for the prompt)
--outdir [dir] - (This is the command for the save directory)
--ddim_steps [steps] - (This is the command for the amount of steps)
--plms - (This is the command to change the sampler to plms)
--n_iter [number] - (how many times to generate images)
--W [width] - (This is the command for setting image width)
--H [height] - (This is the command for setting image height)
--n_samples [number] - (Number of images to generate at the same time)
--scale [scale] - (How close image should match prompt)
--from-file - (This is the command to load prompts from a file)
--seed - (This is the command for setting seed)

davidospina
Автор

Dude, I love the fact that you started with the requirements and explained everything so well.

onelios
Автор

This is so far the best tutorial on installing Stable Diffusion and the ONLY one teaching how to install the GUI locally. Great video!

dDesirie
Автор

WARNING! If you haven't installed Git you will get errors when setting up the environment in this video. ThePrincipalComponent has a good tutorial on it.

eleganto-interpolations
Автор

Oh how fun, I can't wait to play with this. Thanks for the great walk through!

Jabrils
Автор

For anyone with Geforce 16xx graphics cards getting GREEN IMAGES, set the precision to full with "--precision full" without the quotes. Also, the CUDA SDK Toolkit seem to help with parallel processing, but I haven't noticed a difference

hoaxygen
Автор

Great stuff. Only issue I had was not having git installed on the machine, but your pinned comment helped me work around the errors. Great work, entertaining, and released so quickly. Channel is severely under-subscribed.

juancornetto
Автор

You need to create a hugging face account and log in to see the license acceptance link forgot to mention in the video also while the images were gening the video started dropping frames😅

tingtingin
Автор

Thank you for helping the community. Ahh, What would we all do without great people like you and all SD team.

alexgouzanov
Автор

in case anyone else runs into "Error: Cannot find command 'git'..." do:
conda install git

dudesworth
Автор

instead of me searching for the tutorial this tutorial found me. thank you for appearing on my recommended greenscreen man.

itsjustaclint
Автор

Finally, a tutorial which is precise, simple and working. you, kind sir!

bosiak
Автор

Thank you so much for this tutorial! I was able to generate images at 512x512, 100 steps, on a "vintage" GTX 970 with just 4G of vram in 4.99 minutes!
Really a-m-a-z-i-n-g

enricciardi
Автор

THANK YOU! For someone who's been trying to make this work for several days, this was the perfect video. Everything is working finally!

Loglestly
Автор

amazing how you made such a high quality video in such a short amount of time!

treudden
Автор

​ @Andros Forever and others who had trouble installing this. You need to have Git for windows installed first. If you don't this will fail. Here's a way to fix and reinstall as @TingTingin suggested. Stay safe out there!

1. Close out of the anaconda prompt window.
2. Make sure git for windows is installed first. (Download it and install.)
3. Open a file explorer window to this C:\Users\Your User Name Here\.conda\envs
4. Delete the folder "ldm".
5. Open up the anaconda prompt again.
6. cd to wherever you downloaded and unzipped your installer files
7. do the prompt again: conda env create -f environment.yaml

You should now see the ldm environment install correctly with the following:
..
done
#
# To activate this environment, use
#
# $ conda activate ldm
#
# To deactivate an active environment, use
#
# $ conda deactivate

LaughterOnWater
Автор

There was an error during "conda" as not finding "git", I was able to install git and update the environment. Thank you for the detailed introduction - you got me up to speed quickly.

RickeyBowers
Автор

Don't say "any gpu" if it's not going to work for half of all the gpus

jakobcantrelle