New method recalibrates Diffusion Transformers for analog compute-in-memory hardware

Researchers have developed a novel method to recalibrate Diffusion Transformers (DiTs) when used with analog compute-in-memory (CIM) hardware. This approach addresses how CIM's inherent nonidealities distort the classifier-free guidance (CFG) signal, which is crucial for generating high-quality images. By adjusting only the CFG scale during sampling, the method effectively restores generation quality, significantly reducing the gap in FID scores caused by CIM noise across various models like PixArt-Sigma, PixArt-alpha, and DiT-XL/2. AI

IMPACT Improves the efficiency and quality of image generation models on emerging analog hardware, potentially lowering costs and increasing accessibility.

RANK_REASON Academic paper detailing a new method for improving AI model performance on specific hardware. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.CV →

• analog compute-in-memory
• Classifier Free Guidance
• Diffusion Transformers
• DiT-XL/2
• PixArt-alpha
• PixArt-Sigma

• paper
• infra

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →