Stable Diffusion is a machine learning model developed by Stability AI to generate digital images from natural language descriptions. The model can be used for different tasks like generating image-to-image translations guided by text prompts and upscaling images.
Unlike competing models like DALL-E, Stable Diffusion is open source and does not artificially limit the images it produces. Stable diffusion was trained on a subset of the LAION-Aesthetics V2 data set. It can run on most consumer hardware equipped with a modest GPU and was hailed by PC World as "the next killer app for your PC".

Since Stable Diffusion is run locally and not in the cloud, as mentioned there is no limit to the number of images that you can produce but in order to use it you will have to get down a little dirty with setting your PC environment for it since it is not really an application, it is a command line text based descriptor that will use python to generate your images, so there is no install nor GUI.
In this guide, we will show you how to both install and run Stable Diffusion on your local PC so you can start producing some cool images all by yourself.
Hardware and software requirements
Make no mistake, Stable Diffusion will not run on a potato PC, in order to harvest the power of AI-generated imagery this is what you will need:
- A GPU with at least 4GB of VRAM
- 10GB of hard disk space
- Python and libraries (Miniconda3 installer will install everything you need)
- The Stable Diffusion files
- Git
- Any OS (Windows, Linux, macOS)
Installing components
For this tutorial, we are covering the installation and running of Stable Diffusion on Windows PC. The steps presented here are presented in a way that installation can be performed on any operating system but precise instructions will be for Windows OS.
GIT
The first thing to do is to install GIT. It is a tool that will let you easily maintain and install repos from the internet. to install it go to: https://git-scm.com/ and click on download. Follow the instructions for your version of the operating system. If you are a developer you are familiar with GIT and if you already have it installed you can skip this step.
One thing that is important when installing GIT locally is to select to use it via the command line (the second option that says "Git from the command line and also from 3rd-party software").
Miniconda3
Now when we have GIT installed, next thing is to use Miniconda3 to install python and all required libraries that are needed. Get the installer at: https://docs.conda.io/en/latest/miniconda.html
Miniconda3 is basically an easy installer so you do not have to install tons of stuff manually from different websites and sources, it is nicely packaged in the installer that will take care of everything.
Stable Diffusion
After the previous two steps, we are ready now to actually install Stable Diffusion. Go to https://huggingface.co/CompVis/stable-diffusion#model-access and install the latest library (as of the writing of this article currently it is stable-diffusion-v1-4-original, the last one on the right), the library is almost 5GB in size so be prepared for big download.
After installing stable diffusion's latest library it is time to update it to the newest version. You can download ZIP from GIT HUB https://github.com/CompVis/stable-diffusion
Once downloaded click on the Windows start button and type in Miniconda3 and click on open. Create a folder and name it how you want on a drive of your choice. For this example, we will install it all in disk C under folder AI_art, follow the instructions below but use your own names and destination instead. Do not close Minicoda3 after typing commands!!!
cd c:/
mkdir AI_art
cd AI_art
Extract GitHub files that you have downloaded into your new folder and get back to Minicoda3 and type the next commands:
cd C:\AI_art\stable-diffusion-main
conda env create -f environment.yaml
conda activate ldm
mkdir models\ldm\stable-diffusion-v1
Let the whole process finish, some files are large and it might take a while. After the whole process is finished and completed, copy the checkpoint file that you have downloaded into: C:\AI_art\stable-diffusion-main\models\ldm\stable-diffusion-v1
After the file is copied rename it to model.ckpt and you are finished.
Running Stable Diffusion
The created environment is needed in order to actually use Stable Diffusion to create images. Each time you want to use it you will have to run it, so go into Miniconda3, and inside it type:
conda activate ldm
cd C:\AI_art\stable-diffusion-main
after we are inside the folder call the script with the parameters:
python scripts/txt2img.py --prompt "TXT DESCRIPTION OF IMAGE THAT YOU WANT TO CREATE" --plms --n_iter 5 --n_samples 1
and that's it, your image is created and it is located in C:\AI_art\stable-diffusion-main\outputs\txt2img-samples\samples
Next-generation of RAM, DDR5 supposed to hit shelves around late summer or fall in 2021if everything goes as planned.
If we are to believe news from Intel, new series of processors based on Alder Lake architecture are about to be released somewhere in October. Alder Lake is the 12th Generation Core of Intel processors and it is meant to have all advantages of the current CPU structure.
I9 is aimed to pack a punch against its competitors with technology including eight high-performance Golden Cove cores (P-cores) with Hyper-Threading and eight energy-efficient Goldmont (E-cores) equipped with a 30MB L3 cache and operating at high clock speeds.
During the benchmark test, the new CPU outperformed all of Intel’s previous Rocket Lake structure processors which is logically expected but the big surprise came from results that indicate better performance than Apple’s new M1 processor even in single-threaded performance which is supposed to be best in that field.
Another surprise was really beating AMD Ryzen 9 in each field. If we take into account that i9 has 16 cores in total and can process 24 threads against Ryzen 9 with 16 Cores and 32 threads this is an impressive feature.
Now truth to be told Alder lake has massive L3 cache and microarchitectural advantages and consumes much more power but the results of performance are high. Also, note that this early CPU is not really a mass production one and maybe some things will still change but early results are very satisfactory.
We will see what will price tag be since in this age of shortage and competition that will also play a big role in its overall success. Speaking of competition I eagerly await AMD’s response to this, in any way it seems that the CPU market is for some turbulence which is always a good thing for the final customer and I welcome it. 