.png){kind=icon}
.png){kind=icon}
Return
DiffLight: Generative Diffusion Model Acceleration With Silicon Photonics
.png){kind=icon}
PRE
AI
DOI：10.1109/MDAT.2025.3647759.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Editor's notes: Diffusion models are a class of generative AI methods that are increasingly used in real-world applications to generate realistic content. This article designs a hardware accelerator using silicon photonics for diffusion models and experimentally demonstrates its efficacy. -Jana Doppa, Washington State University, USA
Keywords:
Diffusion models
Tuning
Optical network units
Optical buffering
Silicon photonics
Computer architecture
Vertical cavity surface emitting lasers
Optical modulation
Vectors
Performance evaluation
Diffusion Models
Stable Diffusion
Silicon Photonics
Optical Computing