Artistic Qr Code Generator

Artistic Qr Code Generator uses Diffuser to invoke the stable diffusion model and ControlNet. It allows you to do the generation from the command line instead of the GUI.

Try out

Try it in CODIDO or

How does it work?

If you are interesting how the Stable Diffusion and Control Net work together, here is some greate articles to start

• Stable Diffusion QR Code 101
• Refining AI Generated QR Code

How to use?

Build

git clone https://github.com/duanxianpi/artistic_qrcode_docker
cd artistic_qrcode_docker
docker image build -t artistic_qrcode .

Usage

docker container run --gpus all artistic_qrcode_docker -h

usage: app.py [-h] [--input INPUT] [--output OUTPUT] [--codido CODIDO] [--text TEXT] [--prompt PROMPT] [--negative_prompt NEGATIVE_PROMPT]
              [--qr_code_style QR_CODE_STYLE] [--control_net CONTROL_NET] [--art_style ART_STYLE] [--noise NOISE] [--CN_Scale CN_SCALE]
              [--CN_Start CN_START] [--CN_End CN_END]

options:
  -h, --help            show this help message and exit
  --input INPUT         input file will be downloaded from AWS
  --output OUTPUT       output will be upload from AWS to
  --codido CODIDO       running on codido
  --text TEXT           The text you want to encode in QR Code.
  --prompt PROMPT       What do you want your QR Code to look like?
  --negative_prompt NEGATIVE_PROMPT
                        What you don't want to appear on the qr code
  --qr_code_style QR_CODE_STYLE
                        What should the basic qr_code look like?
  --control_net CONTROL_NET
                        Which QR Code ControlNet to use
  --art_style ART_STYLE
                        Two art styles to choose from.
  --noise NOISE         Whether to generate noise in the margins, which would be more favorable for qrcode to hide into the drawing. "true" or
                        "false"
  --CN_Scale CN_SCALE   The weight of the ControlNet. The higher the weight, the more the output will be affected
  --CN_Start CN_START   The percentage of the generation process when the ControlNet starts to take effect.
  --CN_End CN_END       The percentage of the generation process when the ControlNet stops taking effect.

Gallery