Learn&Drop dynamically drops CNN layers during training based on change and continuation scores to reduce forward-pass FLOPs by up to 83% and halve training time on VGG and ResNet models with little accuracy loss.
Efficient and effec- tive training of sparse recurrent neural net- works
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Learn&Drop: Fast Learning of CNNs based on Layer Dropping
Learn&Drop dynamically drops CNN layers during training based on change and continuation scores to reduce forward-pass FLOPs by up to 83% and halve training time on VGG and ResNet models with little accuracy loss.