CoNNect is a differentiable connectivity regularizer that encourages sparse but connected networks, improving structural pruning accuracy when added to DepGraph and LLM-pruner.
Structured pruning of deep convolutional neural networks
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
-
CoNNect: Connectivity-Based Regularization for Structural Pruning
CoNNect is a differentiable connectivity regularizer that encourages sparse but connected networks, improving structural pruning accuracy when added to DepGraph and LLM-pruner.