distributed-optimizer

Tag

Cards List
#distributed-optimizer

SCAPE: Accurate and Efficient LLM Training with Extreme Sparse Communication

arXiv cs.LG · 2d ago Cached

SCAPE is a communication-efficient distributed optimizer that leverages first-moment statistics to enable extreme sparsification for LLM training, preserving accuracy while reducing wall-clock time by up to 43.3%.

0 favorites 0 likes
← Back to home

Submit Feedback