Zhuang Liu

It would be nice to support [ConvNeXt](https://arxiv.org/abs/2201.03545) backbones. It is a very simple model that is purely convolutional. They can serve as a drop-in replacement for ResNet or Swin Transformer...

上面train with sparsity会存下模型

@CF2220160244 Could you share the code (for pruning and training from scratch) and concrete results, so we can look into possible reasons?

@luluvi Sorry but we don't have much experience on mobilenet pruning.

Sorry for the late reply. We calculated FLOPs using https://github.com/apaszke/torch-opCounter

In our models, the residual branch is BN-RELU-CONV-BN-RELU-CONV-BN-RELU-CONV. In the addition, all features from the identity mapping and the last CONV in residual branch are kept. So the main branch...

What I mean by "main branch" is the identity shortcut throughout the network, so there are no BN layers in main branch. Whenever there is an BN, we can do...

Hi @youngfly11, thanks for your interests. DenseNet's code is a little different than VGG's. Unfortunately I am busy with other things now, so I will probably release the code when...

In case you're still interested, we've released our Pytorch implementation here https://github.com/Eric-mingjie/network-slimming, which supports ResNet and DenseNet.

In my experiment, masking out bias did not seem to change accuracy much. I thought this was because if γ is zero, then the output of that channel is the...