Course Overview

Please note: This class is currently being edited into small clips and will replace the raw footage that is currently available for viewing.

Class Description:

This course will dive primarily into two AI technologies that can be utilized to powerful effect in conjunction with procedural workflows: Diffusion Models, such as Stable Diffusion, and Large Language Models, such as Chat GPT. Students will learn how to set up AI workflows in Houdini using existing tools, while also learning the skills and building the confidence to be able modify and build on top of these tools themselves.

Learning Outcomes:

We will cover four projects in 2 sessions:

  • Build a tool that allows the user to write a description, and have Chat GPT output the description to a color ramp parameter (“A rainbow” -> ramp parameter with the colors of the rainbow).
  • Build a tool that allows the user to write a description, and have Chat GPT output the coordinates for a curve, as well as modify existing curve coordinates
  • Build a texture generation workflow using Stable Diffusion and PDG
  • Build a diffusion render system using Latent Consistency Models and ControlNet

By the end of the course, students should have an in-depth understanding of how to utilize cutting-edge AI tools inside of Houdini and the confidence and inspiration to empower their own procedural workflows with AI tools.

Course curriculum

    1. Files

    2. Discord

    1. Session 1 Outline

    2. Session 1 Full Raw Video

    1. Session 2 Outline

    2. Session 2 Full Raw Video

    3. Submit your work

AI for Houdini Artists

  • $50.00
  • 4 hours of video content
Session 1

Large Language Models

This class will start by teaching students how to empower their Houdini tools through text prompting, by learning how to use the Chat GPT API from within Houdini. We will then dive further into this topic, feeding Chat GPT data from Houdini and allowing for the user to modify existing data through text prompting.

  • How to install custom Python libraries in Houdini
  • Setting up the Chat GPT API
  • Chat GPT color ramp tool
  • Chat GPT curve tool
Session 2

Diffusion Models

This class will start by teaching students how to use PDG to build AI workflows. We will use Stable Houdini, an existing Stable Diffusion toolset for Houdini, as well as other AI libraries, to build a texture generation pipeline. From there students will learn how to speed up image generation using Latent Consistency Models and how to control image generation using ControlNet and their own 3D scenes.

  • PDG workflows
  • Stable Houdini
  • Stable Diffusion/Automatic1111
  • Texture generation pipeline
  • External AI libraries (possibly)
  • LCM renders with ControlNet

Instructor

Ryan Gold

Machine Learning and Procedural Engineer

Ryan Gold is a Technical Artist and Machine Learning and Procedural Engineer. He is also the original creator of the Tree Generation Toolset for SideFX Labs. He recently finished an MSc in Artificial Intelligence at the University of St Andrews, researching 3D generative AI and how synthetic data and procedural tools can be best utilized for deep learning, and is now working as a Machine Learning Engineer at the intersection of AI and computer graphics. He has 5 years of experience as a Procedural Technical Artist, building art-directable tools and automatic procedural solutions using Houdini and various real-time engines. He has work experience as a Front-End Developer, Database Engineer, Cloud Engineer, and 3D Modeling and Texture Artist.

LinkedIn Website

WHAT YOU NEED TO TAKE THIS COURSE

  1. Houdini 19.5 or newer (Class will be taught with H20)
  2. Computer (Please see SideFX system requirements)
    1. https://www.sidefx.com/Support/system-requirements/
    2. A second Monitor is recommended, but not necessary
  3. Houdini (Apprentice License is free)

ADDITIONAL INFORMATION

Subscribe to our newsletter.
Add the full Houdini.School schedule to your calendar:

REFUND POLICY:


Multi-Session Courses
Students may request a refund up to 1 day before the start of the course. Students may also withdraw from Multi-Session Courses at any time and are entitled to a pro-rated refund. The withdrawal date must be 1 day before the next class they intend to drop.

On-Demand Courses
All on-demand courses are non-refundable.

Refund a Class
Please send a request to get a refund via email to [email protected].
Your written request to drop any or all of your classes must include:
  1. Student’s full name
  2. Name of the course(s) being dropped

See you in Class!