kaijieshi7/Dynamic-convolution-Pytorch

Issues

• 预训练权重不能加载

  #14 opened 3 years ago by mei123456789
  2

• 无法训练

  #27 opened 9 months ago by si44444444
  3

• what dose K mean？

  #26 opened 9 months ago by AlgerRobben
  0

• 请问如何正确运行_错误是model未定义

  #25 opened a year ago by fxYOLO
  5

• 想请教一下attention2d里的temperature是干嘛的？

  #4 opened 4 years ago by rainylt
  2

• 请问大佬，有没有预训练模型呀？

  #13 opened 3 years ago by ma3252788
  4

• 网络训练

  #20 opened 2 years ago by fenghao195
  1

• 救急救急救急！！！ 调试时发现每个卷积核的权重都是0.25，想问这是什么原因呢？怎么解决呢？

  #22 opened 2 years ago by cvor1
  3

• how to use 1D dynamic convolution

  #24 opened a year ago by Khaled-Mohammed-Abdelgaber
  0

• A Question in Batch Size and torch.Tensor.expand

  #23 opened a year ago by TerenceLiu98
  1

• 当stride 为2的时候咋办啊

  #21 opened 2 years ago by oszn
  1

• Temperature annealing 在使用时如何实现？如何在前10个epoch使用？

  #19 opened 2 years ago by WINGS-999
  1

• 替换mobilenetv2的分组卷积报错

  #7 opened 4 years ago by H-WU
  2

• Dynamic conv3d中的参数初始化

  #18 opened 2 years ago by Aurora-Li11
  2

• 关于Temperature的问题想向您请教，期盼回复

  #16 opened 3 years ago by Ljiaqii
  2

• NotImplementedError: Dilation > 1 not supported in BasicBlock

  #15 opened 3 years ago by yyaoshuai
  0

• 请问一下，这里的attention2d,是什么类型的attention机制呢？

  #1 opened 4 years ago by shuyuandeqipa
  2

• RuntimeError: view size is not compatible with input tensor's size and stride (at least one dimension spans across two contiguous subspaces). Use .reshape(...)

  #12 opened 3 years ago by khawar-islam
  2

• 请问一下，动态权重如何在移动端集成呢？

  #11 opened 3 years ago by yuqiu1233
  2

• 想请教以下用3D动态卷积的时候报错问题

  #5 opened 4 years ago by November666
  9

• No accuracy boost with default settings

  #10 opened 4 years ago by agnonchik
  1

• 感谢复现

  #9 opened 4 years ago by Yuzuki-N
  1

• 关于dynamic convolution使用在时间序列上的问题

  #6 opened 4 years ago by LanceJerry
  1

• The initialization of bias in Dynamic_convNd is undefined

  #8 opened 4 years ago by kjohew
  1

• 请问你实现的动态卷积速度如何

  #3 opened 4 years ago by haoopan
  3

• There is no Dynamic_conv1d in repo, right ?

  #2 opened 4 years ago by xiaoyangnihao
  2