AI Artistry with Krita: A Deep Dive into Various Models and LoRA Implementation
By Ms. Kayla Ng, Executive Officer, Digital Scholarship and Data Analytics, HKU Libraries
Abstract
Discover the possibilities of AI-driven image generation through this demonstration session featuring Krita AI Diffusion ( https://kritaaidiffusion.com/ ). Observe how image generation models work and explore LoRA technology’s role in model customization. Through detailed demonstrations, participants will see various techniques, including image generation, inpainting for artwork modification, and outpainting for canvas expansion. This session offers valuable insights for artists, creative professionals, and those interested in understanding how AI can enhance creative workflows. Perfect for those curious about AI’s role in digital art, even if you are not ready to implement it yet.
Learning Outcomes:
After completing the session, learners will be able to
- Understand image generation model fundamentals
- Learn about LoRA technology for model customization
Required Hardware (for reference only):
Graphics processing options:
- NVIDIA GPUs: CUDA support (6GB+ VRAM recommended)
- AMD GPUs:
- Windows: DirectML support
- Linux: ROCm support
- Apple Silicon: M1/M2 via MPS
- CPU: Supported but significantly slower
About the Speaker
Ms. Kayla Ng
Executive Officer, Digital Scholarship and Data Analytics, HKU Libraries
Ms. Kayla Ng is the Executive Officer for Digital Scholarship and Data Analytics. Her responsibilities include supporting digital initiatives and data analytics projects through data cleansing, metadata formatting, and digitization tasks. With expertise in Python, AI, and project management, she works closely with colleagues to enhance the library’s resources and contribute to academic endeavors.
Floor Plan of the Venue
