primeqa.calibration.train_confidence_calibrator.DataTrainingArguments#
- class primeqa.calibration.train_confidence_calibrator.DataTrainingArguments(dataset_name: str = 'tydiqa', dataset_config_name: str = 'primary_task', overwrite_cache: bool = False, relative_confidence_train_size: float = 0.1, confidence_dataset_dir: Optional[str] = None, preprocessing_num_workers: Optional[int] = None, max_seq_length: Optional[int] = None, max_q_char_len: int = 128, single_context_multiple_passages: bool = False, max_contexts: Optional[int] = None, max_train_samples: Optional[int] = None, max_eval_samples: Optional[int] = None, doc_stride: int = 128, n_best_size: int = 20, n_best_logits: int = 20, max_answer_length: int = 32, negative_sampling_prob_when_has_answer: float = 0.01, negative_sampling_prob_when_no_answer: float = 0.04)#
Bases:
objectArguments pertaining to what data we are going to input our model for training and eval.
Methods
Attributes
confidence_dataset_dirdataset_config_namedataset_namedoc_stridemax_answer_lengthmax_contextsmax_eval_samplesmax_q_char_lenmax_seq_lengthmax_train_samplesn_best_logitsn_best_sizenegative_sampling_prob_when_has_answernegative_sampling_prob_when_no_answeroverwrite_cachepreprocessing_num_workersrelative_confidence_train_sizesingle_context_multiple_passages