Adobe Firefly's Generative Fill and ControlNet Extension

Introduction

• Topic: Utilizing Adobe Firefly's Generative Fill feature with the ControlNet extension.
• Objective: Guide on integrating and using the new feature in Automatic1111.

Prerequisites

• Update ControlNet extension to the latest version.
• Restart Automatic1111 web UI.
• Download the necessary inpaint model.
  • Download link provided in the tutorial.
  • Place the model in the sd-webui-controlnet models folder.

Setting Up

1. Open Automatic1111 Web UI

   • Use Realistic Vision version 20 for experimentation.
   • Ensure ControlNet version 1.1.208 is displayed.

2. Configuring Settings

   • Quick Settings List: Configuration
     • Go to Settings tab.
     • Locate Quick Settings List to display configurations like VAE file.
     • Save settings and restart the web UI.

Using ControlNet with Generative Fill

1. Select a Real Image for Experimentation

   • Example image used in the tutorial.

2. Masking the Image

   • Define the area for generative fill.

3. ControlNet Configuration

   • Selection: Predefined selection using inpaint.
   • Preprocessor & Model: Ensure inpaint preprocessor and the correct model are selected.
     • Refresh if the model isn’t displayed correctly.
   • Settings: Enable ControlNet
     • Pixel perfect option.
     • Enable ControlNet.
     • Set image’s original resolution (e.g., 1280x1024).
   • Generate Image: Hit generate to apply the generative fill.

Results

• Comparison of original image vs. generative fill result.
  • The generative fill feature is impressive and entirely free to use.
  • Potential to avoid costly Adobe subscriptions.

Additional Tips

• Experiment with different sampling methods.
• Apply high-resolution fixes and observe the results.

Tutorials & Community

• Installation Videos: Links provided to tutorials on installing ControlNet and Automatic1111 web UI.
• Community Engagement:
  • Like, subscribe, comment, and share.
  • Join the Discord server for latest updates and discussions.

Additional Resources

• Related Videos: DeepFake tutorial video.
• Support: Join the channel on YouTube or Patreon for more content.

Closing

• Join the growing Discord community.
• Stay updated with the latest in Stable Diffusion and AI technologies.