LoRA Rank Scaling in Cross-Attention Layers and Its Impact on Wan2.1 I2V-14B Inference Efficiency

This report synthesises findings from 15 peer-reviewed papers addressing the following research question: How does the LoRA rank scaling in cross-attention layers affect the inference efficiency (in tokens/second) of Wan2.1 I2V-14B compared to full fine-tuning on downstream video synthesis tasks. With the breakthroughs in deep learning, the recent years have witnessed a booming of artificial intelligence (AI) applications and services, spanning from personal assistant to recommendation systems to video/audio surveillance. More recently, with the proliferation of mobile. 6 claims were extracted from source literature; 6 were independently verified against retrieved documents. An automated multi-reviewer quality assessment produced a score of 8.5/10. This report is a machine-generated literature synthesis and does not constitute original research.

Research goal: How does the LoRA rank scaling in cross-attention layers affect the inference efficiency (in tokens/second) of Wan2.1 I2V-14B compared to full fine-tuning on downstream video synthesis tasks?

Autonomous literature synthesis. Automated review score: 8.5/10. Full text and citation available at Assignee Research.