megvii-model/WeightNet

Issues

• in the WeightNet_DW，is that "x_w = x_w.reshape(-1, 1, 1, self.ksize, self.ksize)"wrong？why the input and output channel is 1?

  #9 opened a year ago by kairenchen123
  0

• 5-D tensors as weight for Conv2d?

  #8 opened 4 years ago by ruiyuanlu
  1

• pytorch version

  #7 opened 4 years ago by czy112
  4

• RetinaNet-WeightNet can not convergence ，with loss = 1.72

  #4 opened 4 years ago by convolutionROC
  1

• error with distributed training

  #6 opened 4 years ago by jmkim0309
  2

• 如何迁移学习自己的数据集

  #5 opened 4 years ago by hehaiyang111
  1

• Should the standard pointwise convolution in the skip path of Resblock (upsample/downsample=True) be kept or also replaced by WeightNet?

  #2 opened 4 years ago by ZhouYuxuanYX
  2

• Why is the bias of wn_fc2 of class WeightNet in weightnet.py set to be false?

  #1 opened 4 years ago by ZhouYuxuanYX
  2