torch_frame.nn.decoder.ExcelFormerDecoder

class ExcelFormerDecoder(in_channels: int, out_channels: int, num_cols: int)[source]

Bases: Decoder

The ExcelFormer decoder introduced in the “ExcelFormer: A Neural Network Surpassing GBDTs on Tabular Data” paper.

Parameters:

  • in_channels (int) – Input channel dimensionality

  • out_channels (int) – Output channel dimensionality

  • num_cols (int) – Number of columns.

reset_parameters() → None[source]

Resets all learnable parameters of the module.

forward(x: Tensor) → Tensor[source]

Transforming x into output predictions.

Parameters:

x (Tensor) – Input column-wise tensor of shape [batch_size, num_cols, in_channels]

Returns:

[batch_size, out_channels].

Return type:

Tensor