pyabsa.augmentation.aug_utils
Module Contents
Functions
perform contextual noise on code, based on replace, insert, delete operations |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
- pyabsa.augmentation.aug_utils.contextual_code_noise_instance(code: str, noise_level: float = 0.15, noise_type: str = 'hybrid', **kwargs) str [source]
perform contextual noise on code, based on replace, insert, delete operations :param code: input code :param noise_level: noise level :param noise_type: noise type, can be {word, char, token} :param kwargs: other arguments :return: augmented instance
- pyabsa.augmentation.aug_utils.contextual_noise_instance(text: str, tokenizer, noise_level: float = 0.15, noise_type: str = 'word', **kwargs)[source]
- Parameters:
text – input text
tokenizer – tokenizer
noise_level – noise level
noise_type – noise type, can be {word, char, token}
kwargs – other arguments
- Returns:
augmented instance
- pyabsa.augmentation.aug_utils.__word_noise_instance(text, tokenizer, noise_level, **kwargs)[source]
- Parameters:
text – input text
tokenizer – tokenizer
noise_level – noise level
kwargs – other arguments
- Returns:
augmented instance
- pyabsa.augmentation.aug_utils.__char_noise_instance(text, tokenizer, noise_level, **kwargs)[source]
- Parameters:
text – input text
tokenizer – tokenizer
noise_level – noise level
kwargs – other arguments
- Returns:
augmented instance
- pyabsa.augmentation.aug_utils.__token_noise_instance(text, tokenizer, noise_level, **kwargs)[source]
- Parameters:
text – input text
tokenizer – tokenizer
noise_level – noise level
kwargs – other arguments
- Returns:
augmented instance
- pyabsa.augmentation.aug_utils.contextual_ids_noise_instance(ids: List[int], tokenizer, noise_level: float = 0.15, noise_type: str = 'mask', **kwargs)[source]
- Parameters:
ids – input ids
tokenizer – tokenizer
noise_level – noise level
noise_type – noise type, can be {word, char, token}
kwargs – other arguments
- Returns:
augmented instance