simrun ❭ modular_reduced_model_inference ❭ strategy ❭ Strategy_spatiotemporalRaisedCosine ❭ _convert_x_static

Strategy_spatiotemporalRaisedCosine._convert_x_static¶

static simrun.modular_reduced_model_inference.strategy.Strategy_spatiotemporalRaisedCosine._convert_x_static(groups, len_z, x)¶

Convert the input array \(\mathbf{x}\) into a dictionary of basis vectors.

Useful for passing the learnable weights to the optimizer as a one-dimensional array, but keeping track of the basis vectors for each group and dimension.

Parameters:¶

• groups (list) – The list of groups.

• len_z (int) – The length of the spatial domain.

• x (array) – The one-dimensional input array.

Returns:¶

A dictionary of basis vectors for each group.

Return type:¶

dict

Example

>>> x
array([0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1.])
>>> groups
['EXC', 'INH']
>>> len_z
3                               # implies len_t is len(x)/2 - len_z = 2
>>> _convert_x_static(groups, len_z, x)
{
    'EXC': (
        array([0.1, 0.2, 0.3]), # spatial
        array([0.4, 0.5])),     # temporal
    'INH': (
        array([0.6, 0.7, 0.8]), # spatial
        array([0.9, 1.]))}      # temporal