ConvAEDecoder

class behavenet.models.aes.ConvAEDecoder(hparams)[source]

Bases: behavenet.models.base.BaseModule

Convolutional decoder.

Methods Summary

build_model()

Construct the decoder.

forward(x[, pool_idx, target_output_size, …])

Process input data.

Methods Documentation

build_model()[source]

Construct the decoder.

forward(x, pool_idx=None, target_output_size=None, dataset=None)[source]

Process input data.

Parameters

  • x (torch.Tensor object) – input data

  • pool_idx (list) – max pooling indices from encoder for unpooling

  • target_output_size (list) – layer-specific output sizes from encoder for unpooling

  • dataset (int) – used with session-specific io layers

Returns

shape (n_input_channels, y_pix, x_pix)

Return type

torch.Tensor