A method that gradually freezes low-importance adapters during PEFT reduces memory usage by about 40-50% with comparable or slightly better accuracy on several benchmarks.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2024 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Not All Adapters Matter: Selective Adapter Freezing for Memory-Efficient Fine-Tuning of Language Models
A method that gradually freezes low-importance adapters during PEFT reduces memory usage by about 40-50% with comparable or slightly better accuracy on several benchmarks.