Introduction to Warpfusion

In this session, you'll learn how to utilize Warpfusion to process video-to-video generations. Warpfusion utilizes Stable Diffusion to generate user customized images for each frame. We will be able to control and customize Stable Diffusion with several tools including ControlNet.

Requirements:

• At least 2GB available on your Google Drive
• Google Colab Pro or Colab Pro+
• At least 16GB of Nvidia GPU memory
• Warpfusion script v15

Directions:

1. Open the Notebook in Colab
2. Change the runtime type to T4 or higher GPU
3. Run parts 1.1-1.3 to install Stable Diffusion dependencies (~ 6 min)
4. Check the box for Skip Install in part 1.3
5. Restart the kernel and run all all of the cells from the beginning