A retrospective of the authors' LoNAS, Shears, and SQFT methods arguing that elastic low-rank adapters and weight-sharing NAS together compress LLMs efficiently.
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
UNVERDICTED 1representative citing papers
citing papers explorer
-
Low-Rank Adapters Meet Neural Architecture Search for LLM Compression
A retrospective of the authors' LoNAS, Shears, and SQFT methods arguing that elastic low-rank adapters and weight-sharing NAS together compress LLMs efficiently.