medcat.utils.relation_extraction.base_component

Module Contents

Classes

BaseComponent_RelationExtraction

Attributes

logger

medcat.utils.relation_extraction.base_component.logger

class medcat.utils.relation_extraction.base_component.BaseComponent_RelationExtraction(tokenizer=BaseTokenizerWrapper_RelationExtraction(), model=None, model_config=None, config=ConfigRelCAT(), task='train', init_model=False)

Parameters:

• tokenizer (medcat.utils.relation_extraction.tokenizer.BaseTokenizerWrapper_RelationExtraction) –

• model (medcat.utils.relation_extraction.models.BaseModel_RelationExtraction) –

• model_config (medcat.utils.relation_extraction.config.BaseConfig_RelationExtraction) –

• config (medcat.config_rel_cat.ConfigRelCAT) –

• task (str) –

• init_model (bool) –

name = 'base_component_rel'

__init__(tokenizer=BaseTokenizerWrapper_RelationExtraction(), model=None, model_config=None, config=ConfigRelCAT(), task='train', init_model=False)

Component that holds the model and everything for RelCAT.

Parameters:

• tokenizer (BaseTokenizerWrapper_RelationExtraction) – The base tokenizer for RelCAT.

• model (BaseModel_RelationExtraction) – The model wrapper.

• model_config (BaseConfig_RelationExtraction) – The model-specific config.

• config (ConfigRelCAT) – The RelCAT config.

• task (str) – The task - used for checkpointing.

• init_model (bool) – Loads default BERT base model, tokenizer, model config. Defaults to False.

save(save_path)

Saves model and its dependencies to specified save_path folder.

The CDB is obviously not saved, it is however necessary to save the tokenizer used.

Parameters:

save_path (str) – folder path in which to save the model & deps.

Return type:

None

classmethod load(pretrained_model_name_or_path='./')

Parameters:

pretrained_model_name_or_path (str) – Path to RelCAT model. Defaults to “./”.

Returns:

BaseComponent_RelationExtraction – component.

Return type:

BaseComponent_RelationExtraction