Functions — github.com/ShannonAI/Entity-Relation-As-Multi-Turn-QA

hub / github.com/ShannonAI/Entity-Relation-As-Multi-Turn-QA / functions

Functions35 in github.com/ShannonAI/Entity-Relation-As-Multi-Turn-QA

⨍ Functions 35 ◇ Types & classes 4

↓ 5 callersFunctionget_question

Args: head_entity: (entity_type,start_idx,end_idx,entity_string) or entity_type

preprocess.py:163

↓ 4 callersFunctionget_inputs

(context, q, tokenizer, title="", max_len=200, ans=[], head_entity=None)

dataloader.py:69

↓ 2 callersFunctionget_score

(gold_set, predict_set)

evaluation.py:7

↓ 2 callersMethodinit_data

(self)

dataloader.py:150

↓ 2 callersFunctionload_t1_data

(dataset_tag, test_path, pretrained_model_path, window_size, overlap, batch_size=10, max_len=512)

dataloader.py:379

↓ 2 callersFunctionpassage_blocks

(txt, window_size, overlap)

preprocess.py:105

↓ 2 callersFunctiontag_decode

(tags, context_mask=None)

evaluation.py:110

↓ 2 callersFunctiontest_evaluation

(model, t1_dataloader, threshold, amp=False)

evaluation.py:16

↓ 1 callersFunctionaligment_ann

(original, newtext, ann_file, offset)

preprocess.py:10

↓ 1 callersFunctionargs_parser

()

train.py:27

↓ 1 callersFunctionblock2qas

Args: ber: (block,entities,relations) dataset_tag: type of dataset title: title corresponding to the passage to which

preprocess.py:178

↓ 1 callersFunctionchar_to_wordpiece

(passage, entities, tokenizer)

preprocess.py:256

↓ 1 callersFunctioneval_t1

Args: predict: [(s1,e1),(s2,e2),(s3,e3),...] gold: (passage_id,(entity_type,start_idx,end_idx,entity_str)) ids: (pass

evaluation.py:59

↓ 1 callersFunctioneval_t2

Args: predict: [(s1,e1),(s2,e2),(s3,e3),...] gold: (passage_id,(head_entity,relation_type,end_entity)) ids: (passage

evaluation.py:80

↓ 1 callersFunctionget_block_er

Get the block level annotation Args: txt: text to be processed, list of token entities: list of (entity_type, start, end,

preprocess.py:115

↓ 1 callersFunctionload_data

(dataset_tag, file_path, batch_size, max_len, pretrained_model_path, dist=False, shuffle=False, threshold=5)

dataloader.py:355

↓ 1 callersFunctionload_t2_data

(t1_dataset, t1_predict, batch_size=10, threshold=5)

dataloader.py:387

↓ 1 callersFunctionprocess

Args: data_dir: data directory output_dir: output directory tokenizer: tokenizer of pretrained model is_test

preprocess.py:271

↓ 1 callersFunctionquery2relation

query -> <entity_type,relation_type,entity_type>

dataloader.py:117

↓ 1 callersFunctionreload_data

(old_dataloader, batch_size, max_len, threshold, local_rank, shuffle=True)

dataloader.py:365

↓ 1 callersFunctionset_seed

(seed)

train.py:20

↓ 1 callersFunctiontrain

(args, train_dataloader)

Args: dataset_tag: type of dataset path: path to training set file tokenizer： tokenizer of pretraine

dataloader.py:132

Method__init__

Args: dataset_tag: type of dataset test_path: path to test set file tokenizer: tokenizer of pretraine

dataloader.py:201

Method__init__

Args: t1_dataset: an instance of T1Dataset t1_predict: predictions of the first turn QA threshold: on

Args: input: （batch,seq_len） attention_mask: (batch,seq_len) token_type_ids: (batch,seq_len)

model.py:16