src.model.deeplearn.layer.torf_reduce_sum_layer

Classes

TORFReduceSumLayer(*args, **kwargs)

class src.model.deeplearn.layer.torf_reduce_sum_layer.TORFReduceSumLayer(*args, **kwargs)
Author:

Alberto M. Esmoris Pena

Reduce a tensor by summation along a given axis, optionally keeping the reduced dimension and adding a small epsilon for numerical stability.

\[f(\mathbf{X}) = \sum_{\text{axis}} \mathbf{X} \;+\; \varepsilon\]
Variables:

  • axis (int) – The axis along which to sum.

  • keepdims (bool) – Whether to keep the reduced dimension.

  • eps (float) – Small constant added after summation (default 0).

__init__(axis=1, keepdims=True, eps=0.0, **kwargs)

See Layer and Layer.__init__().

call(inputs, training=False, mask=False)

Sum the input tensor along the configured axis.

Parameters:

inputs – The input tensor.

Returns:

The reduced tensor.

get_config()

Return necessary data to deserialize the layer.

classmethod from_config(config)

Use given config data to deserialize the layer.