tensorf 정리

TensorVMSplit(
  (density_plane): ParameterList(
      (0): Parameter containing: [torch.float32 of size 1x16x100x100 (GPU 0)]
      (1): Parameter containing: [torch.float32 of size 1x16x100x100 (GPU 0)]
      (2): Parameter containing: [torch.float32 of size 1x16x100x100 (GPU 0)]
  )
  (density_line): ParameterList(
      (0): Parameter containing: [torch.float32 of size 1x16x100x1 (GPU 0)]
      (1): Parameter containing: [torch.float32 of size 1x16x100x1 (GPU 0)]
      (2): Parameter containing: [torch.float32 of size 1x16x100x1 (GPU 0)]
  )
  (app_plane): ParameterList(
      (0): Parameter containing: [torch.float32 of size 1x48x100x100 (GPU 0)]
      (1): Parameter containing: [torch.float32 of size 1x48x100x100 (GPU 0)]
      (2): Parameter containing: [torch.float32 of size 1x48x100x100 (GPU 0)]
  )
  (app_line): ParameterList(
      (0): Parameter containing: [torch.float32 of size 1x48x100x1 (GPU 0)]
      (1): Parameter containing: [torch.float32 of size 1x48x100x1 (GPU 0)]
      (2): Parameter containing: [torch.float32 of size 1x48x100x1 (GPU 0)]
  )
  (basis_mat): Linear(in_features=144, out_features=27, bias=False)
  (renderModule): MLPRender_Fea(
    (mlp): Sequential(
      (0): Linear(in_features=150, out_features=128, bias=True)
      (1): ReLU(inplace=True)
      (2): Linear(in_features=128, out_features=128, bias=True)
      (3): ReLU(inplace=True)
      (4): Linear(in_features=128, out_features=3, bias=True)
    )
  )
)

renderModule에서 nn.Linear의 첫번쨰 intput 같은 경우에는,
2 $\times$ viewpe $\times$ 3 +
2 $\times$ feape $\times$ P(=27) +
3 + P(=27)
의 input channel로 nn.Linear를 만든다.