A block-level pruning criterion estimates channel importance from activation magnitudes and weight norms in a single forward pass.
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
-
LLM-BIP: Structured Pruning for Large Language Models with Block-Wise Forward Importance Propagation
A block-level pruning criterion estimates channel importance from activation magnitudes and weight norms in a single forward pass.