Skip to content

Scaden

ScadenModule

Bases: Module

ScadenModule is a simple implementation of the Scaden model from the original implementation. paper: https://www.nature.com/articles/s41467-022-34550-9

This module is a combination of three ScadenBaseModule with different parameters.

Attributes:

Name Type Description
model_256 ScadenBaseModule

ScadenBaseModule with 256 hidden units.
model_512 ScadenBaseModule

ScadenBaseModule with 512 hidden units.
model_1024 ScadenBaseModule

ScadenBaseModule with 1024 hidden

Methods:

Name Description
forward

Forward pass of the model.
Source code in src/pydeconv/model/nn/scaden.py 
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
@requires_torch
class ScadenModule(nn.Module):
    """ScadenModule is a simple implementation of the Scaden model from the original implementation.
    paper: https://www.nature.com/articles/s41467-022-34550-9

    This module is a combination of three ScadenBaseModule with different parameters.

    Attributes
    ----------
    model_256 : ScadenBaseModule
        ScadenBaseModule with 256 hidden units.
    model_512 : ScadenBaseModule
        ScadenBaseModule with 512 hidden units.
    model_1024 : ScadenBaseModule
        ScadenBaseModule with 1024 hidden

    Methods
    -------
    forward(x)
        Forward pass of the model.
    """

    def __init__(self, input_dim: int, output_dim: int, model_params_dict: dict):
        """ScadenModule is a simple implementation of the Scaden model from the original implementation.

        Parameters
        ----------
        input_dim : int
            Input dimension.
        output_dim : int
            Output dimension.
        model_params_dict : dict
            model parameters.
        """
        super().__init__()

        # Initialize the models with different parameters
        self.model_256 = _ScadenBaseModule(input_dim=input_dim, output_dim=output_dim, **model_params_dict["model_256"])
        self.model_512 = _ScadenBaseModule(input_dim=input_dim, output_dim=output_dim, **model_params_dict["model_512"])
        self.model_1024 = _ScadenBaseModule(
            input_dim=input_dim, output_dim=output_dim, **model_params_dict["model_1024"]
        )

    def forward(self, x):
        """Forward pass of the model.

        Parameters
        ----------
        x : torch.Tensor
            Input tensor.

        Returns
        -------
        torch.Tensor
            Output
        """
        # Get predictions from each model
        output_1 = self.model_256(x)
        output_2 = self.model_512(x)
        output_3 = self.model_1024(x)

        # Combine the outputs (e.g., averaging)
        output = (output_1 + output_2 + output_3) / 3
        return output

__init__(input_dim, output_dim, model_params_dict)

ScadenModule is a simple implementation of the Scaden model from the original implementation.

Parameters:

Name Type Description Default
input_dim int

Input dimension.

 required
output_dim int

Output dimension.

 required
model_params_dict dict

model parameters.

 required
Source code in src/pydeconv/model/nn/scaden.py 
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
def __init__(self, input_dim: int, output_dim: int, model_params_dict: dict):
    """ScadenModule is a simple implementation of the Scaden model from the original implementation.

    Parameters
    ----------
    input_dim : int
        Input dimension.
    output_dim : int
        Output dimension.
    model_params_dict : dict
        model parameters.
    """
    super().__init__()

    # Initialize the models with different parameters
    self.model_256 = _ScadenBaseModule(input_dim=input_dim, output_dim=output_dim, **model_params_dict["model_256"])
    self.model_512 = _ScadenBaseModule(input_dim=input_dim, output_dim=output_dim, **model_params_dict["model_512"])
    self.model_1024 = _ScadenBaseModule(
        input_dim=input_dim, output_dim=output_dim, **model_params_dict["model_1024"]
    )

forward(x)

Forward pass of the model.

Parameters:

Name Type Description Default
x Tensor

Input tensor.

 required

Returns:

Type Description
Tensor

Output
Source code in src/pydeconv/model/nn/scaden.py 
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
def forward(self, x):
    """Forward pass of the model.

    Parameters
    ----------
    x : torch.Tensor
        Input tensor.

    Returns
    -------
    torch.Tensor
        Output
    """
    # Get predictions from each model
    output_1 = self.model_256(x)
    output_2 = self.model_512(x)
    output_3 = self.model_1024(x)

    # Combine the outputs (e.g., averaging)
    output = (output_1 + output_2 + output_3) / 3
    return output