wofmanaf/SA-Net

Issues

• 请问为什么GN可以获取空间注意力

  #26 opened 2 years ago by DragonPengg
  1

• SAnet是否可以替换其他的域自适应网络呢？

  #27 opened a year ago by xiaoxusanheyi
  0

• 你好，请问sa模块是否可以作为即插即用的子模块，用在特征图之后的其他任务分支上？

  #14 opened 3 years ago by n1-k0
  5

• 参数里的groups=64是怎么来的啊，我可以改成其他的吗

  #25 opened 2 years ago by taotao6
  0

• 热力图问题

  #24 opened 2 years ago by asd646942825
  1

• the password of baiduwangpan

  #23 opened 2 years ago by Zhuoyi416
  0

• Getting error

  #16 opened 3 years ago by DeviSPhd
  2

• ValueError: not enough values to unpack (expected 2, got 1)

  #19 opened 3 years ago by fanghua2021
  3

• 作者你好，提出的SA可以在一维时间序列中使用吗？

  #22 opened 2 years ago by dadapongI6
  0

• 作者你好，请问使用sa时，如果我有50个特征图，那么sa的groups为多少比较好呢

  #21 opened 3 years ago by jjjjjjjjj58
  0

• 请问有pth格式的sa_resnet50的预训练权重吗

  #20 opened 3 years ago by XhHello
  0

• 你好，请问代码实现全连接层为什么用的nn.parameter?

  #18 opened 3 years ago by Qcandy
  0

• 中文名称

  #17 opened 3 years ago by zzybj
  0

• The position of SA in Resnet50 in the model

  #15 opened 3 years ago by syh-hue
  2

• 加入SA层，检测任务效果变差

  #4 opened 4 years ago by ggyybb
  7

• SA模块添加位置的疑问

  #6 opened 4 years ago by Suncheng2019
  3

• 加载pretrain模型问题

  #13 opened 3 years ago by cxy1996
  0

• 你好，通道和空间注意力参数大小都是c/2Gx1x1大小，那么空间注意力岂不是相当于还是通道注意力

  #12 opened 3 years ago by swjtulinxi
  1

• shuffle的作用

  #11 opened 3 years ago by yzk-lab
  3

• spatial attention的疑问

  #8 opened 3 years ago by XinYu-Andy
  2

• 代码的疑问

  #10 opened 4 years ago by xiaoliaobin
  0

• 预训练权重

  #9 opened 4 years ago by cymdhx
  1

• 预训练模型格式

  #7 opened 4 years ago by yzk-lab
  5

• 请问预训练模型是否能发布？

  #2 opened 4 years ago by tengtengyou
  2

• 关于代码的疑惑

  #5 opened 4 years ago by xyl-507
  2

• FIG.4

  #3 opened 4 years ago by LCCmango
  1

• 关于sa_layers

  #1 opened 4 years ago by 3AllenIverson
  1