DReSS adds a data-driven regularization step before structured pruning to keep information in the unpruned channels, improving pruned LLM performance.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
DReSS: Data-driven Regularized Structured Streamlining for Large Language Models
DReSS adds a data-driven regularization step before structured pruning to keep information in the unpruned channels, improving pruned LLM performance.