This repository provides a comprehensive guide on how to create QR codes Art using Stable Diffusion.

• Prerequisites
• Open WebUI
• QR Code Generation
• Others

1. Download this repository and open AI-QRCode.ipynb with Google Colab.
2. Open Resources and set the Hardware accelerator to GPU.

Follow the steps in the code to open the Public URL.

Navigate to Settings and ControlNet. Make sure Multi ControlNet: Max models amount is set to more than 2 (default is 3), then click Apply settings and Reload UI.

Follow the steps below to generate your QR code Art:

1. Open ControlNet -> ControlNet Unit 0 and upload your QRCode.
2. Adjust your settings as follows:
   • set the preprocessor to [invert] if your image has a white background and black lines to Enable
   • Set Preprocessor to inpaint_global_harmonious.
   • Set Model to control_v1p_sd15_brightness [5f6aa6ed].
   • Set Control Weight to 0.35.
3. Open ControlNet -> ControlNet Unit 1 and upload your QRCode, then adjust your settings as follows:
   • set the preprocessor to [invert] if your image has a white background and black lines to Enable
   • Set Preprocessor to inpaint_global_harmonious.
   • Set Model to control_v11f1e_sd15_tile [a371b31b].
   • Set Control Weight to 0.55.
   • Set Starting Control Step to 0.35.
   • Set Ending Control Step to 0.75.
4. Prompt your QR Code image to be generated. For instance:
   • Prompt: masterpiece, best quality, Japanese house, tree, and city background
   • Nagative Prompt: UnrealisticDream, FastNegativeEmbedding

The style of the generated image heavily depends on the Stable Diffusion checkpoint. You can change the checkpoint to generate a QR code image with a different style.

You can add more checkpoints by modifying these lines of code:

To download checkpoints, run curl -Lo filename.safetensors downloadLink and move the downloaded file into the /content/microsoftexcel/models/Stable-diffusion folder.

This repo is inspired by sd-misc-colab and Rowan Cheung's Twitter post.