Polygence Scholar2025
Felix Deng
Class of 2026washington
About
Projects
- "How can lightweight diffusion algorithms be leveraged in generative art to achieve a balance between preserving fine details and minimal operation cost?" with mentor Vardhan (Working project)
Felix's Symposium Presentation
Project Portfolio
How can lightweight diffusion algorithms be leveraged in generative art to achieve a balance between preserving fine details and minimal operation cost?
Started Feb. 14, 2023
Abstract or project description
In this research paper, we will try to optimize diffusion models for fine details. Often diffusion models produce extremely detailed results. However, those results require a significant amount of computing power and data in the form of images. This makes diffusion models extremely expensive, and thus harder to use. Through this research paper, we will find optimizations for diffusion models to improve runtime and require less data.