PGB prunes BERT in a single pass by permuting weight matrices into block-diagonal groups of important weights and removing all other weights, matching or beating iterative baselines on GLUE and SQuAD.
4171-–4186
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
PGB: One-Shot Pruning for BERT via Weight Grouping and Permutation
PGB prunes BERT in a single pass by permuting weight matrices into block-diagonal groups of important weights and removing all other weights, matching or beating iterative baselines on GLUE and SQuAD.