
Introduction:
Recently, a captivating Reddit post caught my attention. It showcased an exquisite collection of artistic QR codes crafted using the incredible power of Stable Diffusion. These remarkable QR codes were generated using a custom trained ControlNet model. The Stable Diffusion community continues to astonish with their creativity, as they have now discovered how to create QR codes without relying on a custom model.
How To make a QR code with Stable Diffusion, you can follow the steps outlined in the tutorial on the webpage you provided. Here is a summary of the process:
- Software: Use the AUTOMATIC1111 Stable Diffusion GUI, which can be used on Google Colab, Windows, or Mac. Make sure you have the ControlNet extension installed.
- Generating QR Code: Start by generating a QR code using a QR Code generator. Make sure the QR code meets the following criteria:
- Use a high fault tolerance setting (30%).
- Have a white margin around the QR code (the quiet zone).
- Use a basic square fill with a black-and-white pattern.
- Avoid generators that introduce a thin white line between black elements.
- Decorate the QR code with Stable Diffusion:
- In the AUTOMATIC1111 WebUI, go to the Img2img page.
- Select the checkpoint model “GhostMix.”
- Enter a prompt and a negative prompt. The prompt should blend naturally with your QR code.
- Upload the QR code to the img2img canvas.
- Set the image-to-image settings as follows:
- Resize mode: Just resize
- Sampling method: DPM++2M Karras
- Sampling step: 50
- Width: 768
- Height: 768
- CFG Scale: 7
- Denoising strength: 0.75
- Upload the QR code to ControlNet’s image canvas.
- Enable ControlNet and set the Control Type to “Tile.”
- Use the preprocessor “tile_resample.”
- Choose the model “control_xxx_tile.”
- Set the Control Weight to 0.87.
- Adjust the Starting Control Step and Ending Control Step as needed.
- Press the “Generate” button to create the QR code artwork.
After following these steps, you can check the generated QR code with your phone’s camera app to ensure it works properly. Remember that not all images will result in a functional QR code, so you may need to generate multiple images and test them to find the successful ones.
Additionally, the tutorial provides some tips and alternative prompts you can use to create different QR code artworks.