cursus.processing.temporal.temporal_mask_processor

Temporal Mask Processor for Temporal Self-Attention Model

This module provides atomic attention mask generation for temporal sequences. Derived from TSA attention masking requirements.

class TemporalMaskProcessor(padding_value=0, output_format='boolean', mask_value=True)[source]

Bases: Processor

Generates attention masks for padded sequences.

Derived from TSA attention masking requirements.

Parameters:

  • mask_value (int | float | bool) – Value indicating valid positions

  • padding_value (int | float) – Value indicating padded positions

  • output_format (str) – ‘boolean’, ‘float’, ‘int’

fit(data)[source]

No fitting required for masking

process(input_data)[source]

Generate attention mask

create_causal_mask(sequence_length)[source]

Create a causal (lower triangular) attention mask.

Parameters:

sequence_length (int) – Length of the sequence

Returns:

Causal attention mask

Return type:

ndarray

create_padding_mask(sequence_lengths, max_length)[source]

Create padding masks for batch of sequences with different lengths.

Parameters:

  • sequence_lengths (List[int]) – List of actual sequence lengths

  • max_length (int) – Maximum sequence length (padded length)

Returns:

Batch of padding masks

Return type:

ndarray

combine_masks(*masks)[source]

Combine multiple masks using logical AND.

Parameters:

*masks (ndarray) – Variable number of mask arrays

Returns:

Combined mask

Return type:

ndarray

get_config()[source]

Return processor configuration