A Kronecker-product binary mask makes the number of groups in each convolution layer learnable end-to-end, giving 0.5 to 1.7 percent top-1 ImageNet gains over ResNet and ResNeXt at comparable cost.
A constructive arbitrary- degree kronecker product decomposition of tensors
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2019 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Differentiable Learning-to-Group Channels via Groupable Convolutional Neural Networks
A Kronecker-product binary mask makes the number of groups in each convolution layer learnable end-to-end, giving 0.5 to 1.7 percent top-1 ImageNet gains over ResNet and ResNeXt at comparable cost.