liuzhuang13 / slimming

Learning Efficient Convolutional Networks through Network Slimming, In ICCV 2017.

liuzhuang13/slimming Issues

Channel number to be pruned must be a pow of 2 ?
Updated 2 years ago1
sparse-training gets contrary BN distribution
Closed 3 years ago5
Is it reasonable to get a threshold for all bn layers?
Updated 4 years ago
Slimming Resnet
Updated 5 years ago10
cifar10 flops higher than cifar100 on DenseNet(40% pruned)
Updated 5 years ago1
在剪枝程序vggprune.py中遇见程序报错
Updated 5 years ago
What's the formulation of calculating flops?
Updated 5 years ago1
Slimming MobileNet
Updated 5 years ago1
Slimming DenseNet
Updated 6 years ago7
Bias masking in BN layers
Updated 6 years ago1
local subgradient = S*torch.sign(weight)
Updated 6 years ago3
Connection between last conv layer and fc layer
Closed 7 years ago2