peal.adaptors.projection_adaptor.ProjectionAdaptorConfig¶

class peal.adaptors.projection_adaptor.ProjectionAdaptorConfig[source]¶

Config of ProjectionAdaptor.

Parameters:

adaptor_type (str) – Fixed to "ProjectionAdaptor"; used by get_adaptor for dispatch.
category (str) – Fixed to "adaptor".
base_model_config (PredictorConfig or dict or str) – Predictor config (or yaml path) of the classifier to repair; its model_path must contain model.cpl.
base_dir (str) – Output directory; created on construction.
data (DataConfig or dict or None) – test_data (if given) replaces the predictor’s data config so the evaluation runs on a different dataset; data is currently unused.
test_data (DataConfig or dict or None) – test_data (if given) replaces the predictor’s data config so the evaluation runs on a different dataset; data is currently unused.
sparse_dictionary (SparseDictionaryConfig or dict or None) – Config passed to get_sparse_dictionary; its get_components() provides the candidate directions.
projected_component_index_list (list of int) – Indices of the dictionary components whose span is projected out.
partition (int) – Which dataloader to evaluate on: 0 train, 1 validation, 2 test.

Methods

Attributes

`adaptor_type`	The type of adaptor that shall be used.
`category`	The category of the config.
`partition`
`projected_component_index_list`
`sparse_dictionary`
`base_model_config`
`base_dir`
`data`
`test_data`

adaptor_type: str = 'ProjectionAdaptor'¶: The type of adaptor that shall be used. This is necessary to know which pydantic class to use when loading from yaml.

category: str = 'adaptor'¶: The category of the config. Can not be changed for adaptor. This is also necessary to identify which pydantic class to use when loading from yaml.