Picture for Chengqi Xu

Chengqi Xu

University of Southern California

Faster Multi-GPU Training with PPLL: A Pipeline Parallelism Framework Leveraging Local Learning

Add code
Nov 19, 2024
Figure 1 for Faster Multi-GPU Training with PPLL: A Pipeline Parallelism Framework Leveraging Local Learning
Figure 2 for Faster Multi-GPU Training with PPLL: A Pipeline Parallelism Framework Leveraging Local Learning
Figure 3 for Faster Multi-GPU Training with PPLL: A Pipeline Parallelism Framework Leveraging Local Learning
Figure 4 for Faster Multi-GPU Training with PPLL: A Pipeline Parallelism Framework Leveraging Local Learning
Viaarxiv icon