Creating images through artificial intelligence has revolutionized the way creativity and visual production is perceived. One of the most prominent tools to achieve this task is Stable Diffusion XL, one of the pioneers in the generation of images by artificial intelligence from simple text instructions that naturally describes what you want to generate, along with platforms such as Dall.e-2 or Midjourney .
In this case Stable Diffusion XL allows any user, even those with computers that do not have powerful GPUs, to access this technology through Google Colab.
Steps to generate images using Stable Diffusion XL’s AI.
The steps to achieve this are as follows:
-Log in to Google Colab: For those who do not own a high performance machine, the option to use Stable Diffusion XL on any computer is through Google Colab. Following this link will take you to a page where you will find the code needed to install and run Stable Diffusion XL.
-Select the execution environment: Once inside Google Colab, access “Runtime Environment” and select “Change Runtime Environment Type”, making sure to choose the GPU option under “Hardware Accelerator”, a choice that will ensure optimal AI execution.
-Execute the first step: After performing the above configuration you can now run the first data cell by clicking on the play button located to the left of this cell. This step will install the necessary dependencies for Stable Diffusion in the virtual machine. A warning may appear when executing the code, in which case select “Run anyway” to continue.
-Execute the second step: Once the installation of the dependencies is complete, proceed to run the second cell by hovering the mouse over this cell and clicking the run button that will appear. The cell will run smoothly and continue to perform the necessary configuration.
How to configure Stable Diffusion XL and generate images.
-Configure Stable Diffusion XL: After executing the second cell a link will appear to access the Stable Diffusion XL configuration interface. Clicking on the link will open a page where the license terms must be accepted. Once accepted, continue with the configuration by selecting the color theme to be used on the computer and specifying that you will be the only user of the application on the computer.
-Select the models: In the next step you will be able to choose the models you want to download. It is important to activate the two Stable Diffusion XL options: the base model and the refined model. Once you have completed this step, you must confirm your choices to start the installation and select the ComfyUI option.
-Choose the model and generate images: After installation, you have to access the Stable Diffusion XL page and select the “OfficialStableDiffusion” model, choosing the “Stable Diffusion XL 1.0 Base” model.
From this point on, you can write a description (in English) of the image you want to create in the “Prompt” box and clicking “Generate Image” will create the image. You can also add “Negative Prompts” to exclude unwanted elements.
The Stable Diffusion XL page offers settings to control image creation, including the option to activate the “Refiner” model to further improve quality, although this tool is not infallible and may have occasional failures. If this happens, you can refresh the page or restart the process in Google Colab by pressing F5.