dudes will be like “cooking something crazy” and then it’s just a website

Quick experiment playing with the new shadow variables in @webflow

Once you’ve hit a balance between functional and engaging, you’re all set! To push it further, drop your output into photoshop beta to outpaint and add some beautiful details

The strength and conditioning scale parameters make the biggest impact outside of the prompt I created this reference chart to understand their influence on each other Figma for HD: https://t.co/EutjTBxnZ6

Be sure to test your output. There’s a decent chance the first version doesn’t scan but no worries, let’s tweak it

With the default settings, click ‘run' Right now, the avg queue time is ~200 seconds but if you duplicate the space and host it yourself it’s closer to 20 seconds

Now add your prompt. This will theme the QR code - I skip the negative prompt unless I’m tweaking the output Prompt Examples: sky view of futuristic city underwater atlantis futuristic cyberpunk cityscape aerial view of ocean waves

First, let’s set the link. Type in the target URL or upload an existing QR code below

We’re using a space called QR Code AI Art Generator Basically generates a QR code then uses a ControlNet pipeline to blend with your prompt https://t.co/kFGI20pjog

Want to make those generative QR codes you’ve seen around? Here’s a 2 minute guide 🤙

If you’re still hitting errors w/ ControlNet, here’s a guide on running it free through Google Colab

Great breakdown of ControlNet from @kanekallaway, thanks for the mention 🤙

A quick reference guide to see how stroke width can impact your ControlNet output image

Experimenting with ControlNet? Try outpainting with Photoshop AI for some incredible results and a huge increase in image quality

I think we crashed the model server, hopefully the errors stop soon

My new favorite tool combo is ControlNet + Photoshop AI I’ll work for free @Nike

...and stack them into a GIF with EZGif. Done! I'll be doing weekly tutorials like this so hit that follow button if you want to see more 🤙

Let's generate a few more... (aerial + canyon, waves, snow, volcano)

Upload your logo trace image into the Hugging Face Lineart model and add a simple description for the prompt. Click 'run' and let the magic happen. There can be a queue so don't worry if it takes 30-60 seconds (prompt: aerial, jungle)

We're using the Lineart model, so export an outline of your logo as a JPEG. Leave some space around the outside to be filled (black and white seems to work best)