标签：shot 05 0.1 gnn affine eps few stats True

训练

分为两个网络：
Embedding层和GNN度量层

EmbeddingOmniglot

omniglot
EmbeddingOmniglot(
  (conv1): Conv2d(1, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
  (bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
  (conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
  (bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
  (conv3): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), bias=False)
  (bn3): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
  (conv4): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), bias=False)
  (bn4): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
  (fc_last): Linear(in_features=576, out_features=64, bias=False)
  (bn_last): BatchNorm1d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)

MetricNN

MetricNN(
  (gnn_obj): GNN_nl_omniglot(
    (layer_w0): Wcompute(
      (conv2d_1): Conv2d(84, 168, kernel_size=(1, 1), stride=(1, 1))
      (bn_1): BatchNorm2d(168, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv2d_2): Conv2d(168, 126, kernel_size=(1, 1), stride=(1, 1))
      (bn_2): BatchNorm2d(126, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv2d_3): Conv2d(126, 84, kernel_size=(1, 1), stride=(1, 1))
      (bn_3): BatchNorm2d(84, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv2d_4): Conv2d(84, 84, kernel_size=(1, 1), stride=(1, 1))
      (bn_4): BatchNorm2d(84, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv2d_last): Conv2d(84, 1, kernel_size=(1, 1), stride=(1, 1))
    )
    (layer_l0): Gconv(
      (fc): Linear(in_features=168, out_features=48, bias=True)
      (bn): BatchNorm1d(48, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
    )
    (layer_w1): Wcompute(
      (conv2d_1): Conv2d(132, 264, kernel_size=(1, 1), stride=(1, 1))
      (bn_1): BatchNorm2d(264, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv2d_2): Conv2d(264, 198, kernel_size=(1, 1), stride=(1, 1))
      (bn_2): BatchNorm2d(198, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv2d_3): Conv2d(198, 132, kernel_size=(1, 1), stride=(1, 1))
      (bn_3): BatchNorm2d(132, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv2d_4): Conv2d(132, 132, kernel_size=(1, 1), stride=(1, 1))
      (bn_4): BatchNorm2d(132, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv2d_last): Conv2d(132, 1, kernel_size=(1, 1), stride=(1, 1))
    )
    (layer_l1): Gconv(
      (fc): Linear(in_features=264, out_features=48, bias=True)
      (bn): BatchNorm1d(48, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
    )
    (w_comp_last): Wcompute(
      (conv2d_1): Conv2d(180, 264, kernel_size=(1, 1), stride=(1, 1))
      (bn_1): BatchNorm2d(264, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (dropout): Dropout(p=0.3, inplace=False)
      (conv2d_2): Conv2d(264, 198, kernel_size=(1, 1), stride=(1, 1))
      (bn_2): BatchNorm2d(198, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv2d_3): Conv2d(198, 132, kernel_size=(1, 1), stride=(1, 1))
      (bn_3): BatchNorm2d(132, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv2d_4): Conv2d(132, 132, kernel_size=(1, 1), stride=(1, 1))
      (bn_4): BatchNorm2d(132, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv2d_last): Conv2d(132, 1, kernel_size=(1, 1), stride=(1, 1))
    )
    (layer_last): Gconv(
      (fc): Linear(in_features=360, out_features=20, bias=True)
      (bn): BatchNorm1d(20, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
    )
  )
)

标签：shot,05,0.1,gnn,affine,eps,few,stats,True
来源： https://blog.csdn.net/nini_k/article/details/118522237

本站声明： 1. iCode9 技术分享网（下文简称本站）提供的所有内容，仅供技术学习、探讨和分享；
2. 关于本站的所有留言、评论、转载及引用，纯属内容发起人的个人观点，与本站观点和立场无关；
3. 关于本站的所有言论和文字，纯属内容发起人的个人观点，与本站观点和立场无关；
4. 本站文章均是网友提供，不完全保证技术分享内容的完整性、准确性、时效性、风险性和版权归属；如您发现该文章侵犯了您的权益，可联系我们第一时间进行删除；
5. 本站为非盈利性的个人网站，所有内容不会用来进行牟利，也不会利用任何形式的广告来间接获益，纯粹是为了广大技术爱好者提供技术内容和技术思想的分享性交流网站。