The Artificial Intelligences that generate images from a text description are trendy. DALL-E 2 is the most popular and is already available to everyone, but it imposes enough limitations so that it is not abused. Also, although this AI gives you 15 monthly credits to generate 15 free photos, it really is a paid application. Isn’t there an AI like DALL-E for free and without limitations? Yes, it is called Stable Diffusion and here we will show you everything you need to know about it.
What is Stable Diffusion and what is it used for?
Stable Diffusion is a tool that generate images from a text description, using Artificial Intelligence to achieve results that resemble what a human would do. It can also be used to color or fade images, as well as to generate a sequence of images guided by a text description.
Unlike DALL-E, Stable Diffusion is completely free and open source, so you can use it without limitations and at no cost. The company that deals with the development of this AI is Stability AI, although the latent diffusion model (a variety of generative neural network) it uses has been developed by researchers at the University of Munich. In addition, Runway, EleutherAI and LAION have also collaborated in the creation of it.
How to use Stable Diffusion to generate free images from text
As stated before, Stable Diffusion is open source, so you can use it for free legally. However, there are two ways to use it: running the Artificial Intelligence on your computer (a modern and powerful GPU is required) or generating the images with cloud services based on Stable Diffusion (it has its limitations). Next, we will explain both methods, starting with the simplest.
Use Stable Diffusion for free and online from these web pages
If you don’t want to complicate your life, it’s best to go to a website that uses Stable Diffusion to generate free images from text. The operation of these websites is the same as that of DALL-E 2: just write the description of the image you want to create and wait for the AI to work its magic. What websites are we talking about? Of these two:
Access | DreamStudio
Access | hugging face
In DreamStudio you will need to sign in with an account to start generating your images. once you do, You will only have to describe the image you want to create (in English) in the bar below and click on Dream. Before doing so, you can change the generation parameters (width, height, scale, number of generated images, etc.) in the right panel. After generating your images, you can tap the download button in the top right corner of the image to save it.
The main limitation of DreamStudio is that it is paid: Each generated image costs a credit (varies according to the size of the image) that is obtained with real money. The good thing is that they give you 200 free credits when you start.
As for Hugging Face, this is a demo version of Stable Diffusion that you can use for free from your browser without installing anything. It does not require creating an account and it is just as easy to use, since you just have to write the text of what you want to generate and click on Generate image. It takes longer to generate the images than DreamStudio due to the fact that it generates four images at once for each description you pass to it. You can change this (and other things) by tapping the Advanced options.
The good thing about Hugging Face is that it is completely free and does not ask for credits to generate images. The bad thing is that their servers are constantly saturated, so more than once you will see a waiting queue and the error message: try again later.
Finally, we recommend trying Text-to-Pokémon, the AI for create unique Pokémon based on Stable Diffusion.
Install Stable Diffusion on your PC to generate images with AI as you want
You can install Stable Diffusion on your computer to use it without depending on web page servers. The best way to do this is with the NMKD program which, we warn you, installs an uncensored, unfiltered version of Stable Diffusionso you must be careful: you and only you are responsible for the content generated with it.
Requirements to install Stable Diffusion on your PC
for now, this program only works with NVIDIA graphics, so if you have AMD or Intel graphics we are sorry to inform you that it will not work. Also, not every PC is capable of running Stable Diffusion. Your computer must have at least the following characteristics:
- Windows 10 or Windows 11 64 bit.
- NVIDIA graphicspreferably RTX, 10th generation or later.
- 16GB RAM.
- 4GB of VRAM (for slow low-memory mode) or 6 GB of VRAM (for fast mode with more features).
- 11 GB of free storage for installation (preferably on an SSD).
By the way, there is a Known issue with GTX 16 series GPUs, which produce a fully green image. Fixed by turning on low memory mode and full accuracy in settings.
How to install Stable Diffusion on PC with NMKD
Do you meet the requirements? Then follow this tutorial to install Stable Diffusion on your NMKD PC:
- Go to this link to download NMKD by pressing the button download now which is down They will ask you to donate something, but you can skip it by clicking on “No thanks, just take me to the downloads”. Currently, the program weighs 1.7 GB.
- Once it’s downloaded, unzip the ZIP file and double click on the EXE file named «stablediffusiongui«.
- press on Install and let the program install itself, like any other program.
- Clever! Look for the program access icon Stable Diffusion GUI and double click on it to open it and start using it.
In the program you will have the same options as in web pages that generate images from text, or even more. In “Prompt» you must place the text description of the image you want to create (in English) and then press the button Generate! to generate it.
If your PC takes a long time or crashes while generating an image, try changing the value of “Amount of Images to Generate» so that the program generates fewer images at a time. Also, consider lowering the resolution of images to reduce the load on the GPU.
