UNet

UNet model.

A UNet encoder, decoder and complete model.

`UNet`

Bases: Module

UNet model.

Adapted for PyTorch from: https://github.com/juglab/n2v/blob/main/n2v/nets/unet_blocks.py.

Parameters:

Name	Type	Description	Default
`conv_dims`	`int`	Number of dimensions of the convolution layers (2 or 3).	required
`num_classes`	`int`	Number of classes to predict, by default 1.	`1`
`in_channels`	`int`	Number of input channels, by default 1.	`1`
`depth`	`int`	Number of downsamplings, by default 3.	`3`
`num_channels_init`	`int`	Number of filters in the first convolution layer, by default 64.	`64`
`use_batch_norm`	`bool`	Whether to use batch normalization, by default True.	`True`
`dropout`	`float`	Dropout probability, by default 0.0.	`0.0`
`pool_kernel`	`int`	Kernel size of the pooling layers, by default 2.	`2`
`residual`	`bool`	Whether to add a residual connection from the input to the output.	`False`
`final_activation`	`Optional[Callable]`	Activation function to use for the last layer, by default None.	`NONE`
`n2v2`	`bool`	Whether to use N2V2 architecture, by default False.	`False`
`independent_channels`	`bool`	Whether to train the channels independently, by default True.	`True`
`**kwargs`	`Any`	Additional keyword arguments, unused.	`{}`

Forward pass.

Parameters:

Name	Type	Description	Default
`x`	`torch.Tensor`	Input tensor.	required

Returns:

Type	Description
`Tensor`	Output of the model.

Bases: Module

Unet decoder pathway.

Parameters:

Name	Type	Description	Default
`conv_dim`	`int`	Number of dimension of the convolution layers, 2 for 2D or 3 for 3D.	required
`depth`	`int`	Number of decoder blocks, by default 3.	`3`
`num_channels_init`	`int`	Number of channels in the first encoder block, by default 64.	`64`
`use_batch_norm`	`bool`	Whether to use batch normalization, by default True.	`True`
`dropout`	`float`	Dropout probability, by default 0.0.	`0.0`
`n2v2`	`bool`	Whether to use N2V2 architecture, by default False.	`False`
`groups`	`int`	Number of blocked connections from input channels to output channels, by default 1.	`1`

Forward pass.

Parameters:

Name	Type	Description	Default
`*features`	`list[torch.Tensor]`	List containing the output of each encoder block(skip connections) and final output of the encoder.	`()`

Returns:

Type	Description
`Tensor`	Output of the decoder.

Bases: Module

Unet encoder pathway.

Parameters:

Name	Type	Description	Default
`conv_dim`	`int`	Number of dimension of the convolution layers, 2 for 2D or 3 for 3D.	required
`in_channels`	`int`	Number of input channels, by default 1.	`1`
`depth`	`int`	Number of encoder blocks, by default 3.	`3`
`num_channels_init`	`int`	Number of channels in the first encoder block, by default 64.	`64`
`use_batch_norm`	`bool`	Whether to use batch normalization, by default True.	`True`
`dropout`	`float`	Dropout probability, by default 0.0.	`0.0`
`pool_kernel`	`int`	Kernel size for the max pooling layers, by default 2.	`2`
`n2v2`	`bool`	Whether to use N2V2 architecture, by default False.	`False`
`groups`	`int`	Number of blocked connections from input channels to output channels, by default 1.	`1`

Forward pass.

Parameters:

Name	Type	Description	Default
`x`	`Tensor`	Input tensor.	required

Returns:

Type	Description
`list[Tensor]`	Output of each encoder block (skip connections) and final output of the encoder.