CellEncoder.CNN

Click here to view source code.

class CNN(in_dim: int, ft_dim: int = 735, hid_channel_ls: list = None, kernel_size_conv: int = 7,
stride_conv: int = 1, padding_conv: int = 0, kernel_size_pool: int = 3, stride_pool: int = 3,
padding_pool: int = 0, batch_norm: bool = True, max_pool: bool = True, flatten: bool = True,
debug: bool = False)

CNN can be used to encode exp, gsva, mut or cnv of cells.

PARAMETERS:

  • in_dim (int) - Dimension of original feature.

  • ft_dim (int, optional) - Dimension of encoded feature. (default: 735)

  • hid_channel_ls (list, optional) - Hidden channels of CNN. (default: None)

    When the value is None, [40, 80, 60] is used.

  • kernel_size_conv (int, optional) - Kernel size of torch.nn.Conv1d layer. (default: 7)

  • stride_conv (int, optional) - Stride of torch.nn.Conv1d layer. (default: 1)

  • padding_conv (int, optional) - Padding of torch.nn.Conv1d layer. (default: 0)

  • kernel_size_pool (int, optional) - Kernel size of torch.nn.MaxPool1d layer. (default: 3)

  • stride_pool (int, optional) - Stride of torch.nn.MaxPool1d layer. (default: 3)

  • padding_pool (int, optional) - Padding of torch.nn.MaxPool1d layer. (default: 0)

  • batch_norm (bool, optional) - Use torch.nn.BatchNorm1d layer or not. (default: True)

  • max_pool (bool, optional) - Use torch.nn.MaxPool1d layer or not. (default: True)

  • flatten (bool, optional) - Use torch.nn.Flatten() layer or not. (default: True)

  • debug (bool, optional) - Print the shape of tensor f or not. (default: False)

SHAPES:

  • input: Original features [batch_size, in_dim]

  • output: Encoded features [batch_size, ft_dim]

forward(f: torch.Tensor)

  • f (torch.Tensor) - The input of CNN.