Functions — github.com/BladeDancer957/CPFD

hub / github.com/BladeDancer957/CPFD / functions

Functions60 in github.com/BladeDancer957/CPFD

⨍ Functions 60 ◇ Types & classes 7

↓ 6 callersFunctionget_label_distribution

(y_lists,label_list, count=False)

src/dataloader.py:397

↓ 6 callersMethodsave_model

save the best model

src/trainer.py:374

↓ 5 callersFunctionread_ner

(datapath, phase, label_list)

src/dataloader.py:233

↓ 4 callersMethodforward_classifier

(self, features)

src/model.py:52

↓ 4 callersFunctionget_default_label_list

(entity_list, schema='BIO')

src/dataloader.py:215

↓ 4 callersMethodset_unseen_labels_to_O

(self, y_train, seen_label_list)

src/dataloader.py:501

↓ 3 callersMethodevaluate

(self, dataloader, each_class=False, entity_order=[], is_plot_hist=False, is_plot_cm=False)

src/trainer.py:275

↓ 3 callersFunctionget_params

()

src/config.py:6

↓ 2 callersMethod__init__

(self, output_dim, params)

src/model.py:18

↓ 2 callersMethodbatch_forward

(self, inputs)

src/trainer.py:43

↓ 2 callersFunctionentropy

Computes the entropy per token. :param probabilities: Tensor of shape (bsz,seq_len,refer_dims). :return: One entropy per token, shape (bsz,se

src/utils.py:26

↓ 2 callersMethodforward_encoder

features features[0]: hidden_states; features[1]: all_hidden_states(embedding layer + 12 layer output hidden, 13*(bsz, seq_l

src/model.py:41

↓ 2 callersMethodget_dataloader

(self, first_N_classes=-1, select_entity_list=[], phase=['train','dev','test'], is_filter_O=None, filter_entit

src/dataloader.py:510

↓ 2 callersMethodload_model

load the checkpoint

src/trainer.py:390

↓ 2 callersFunctionplot_embedding

Plot the feature X Params: - X: a feature matrix has dims (num_samples, hidden_dims) - Y: a label list has d

src/utils.py:292

↓ 1 callersMethodbatch_backward

(self)

src/trainer.py:267

↓ 1 callersMethodbatch_loss

Cross-Entropy Loss

src/trainer.py:51

↓ 1 callersMethodbatch_loss_cpfd

(self, labels)

src/trainer.py:116

↓ 1 callersMethodbefore

(self, train_loader)

src/trainer.py:101

↓ 1 callersMethodcalculate_sample_weight

(self, labels)

src/trainer.py:104

↓ 1 callersFunctioncompute_class_feature_center

Get features and targets Params: - dataloader: torch.utils.data.DataLoader - feature_model: a model returns

src/utils.py:79

↓ 1 callersFunctioncompute_feature_by_dataloader

Compute the feature of dataloader{(X, Y)}, X has dims (num_sentences, num_words) Params: - dataloader: torch.utils.data.

src/utils.py:118

↓ 1 callersFunctioncreate_logger

Create logger for the experiment Params: - filepath: the path which the log file is saved

src/utils.py:571

↓ 1 callersFunctionentity_sampler

(inputs_train, y_train, label_list, unbalanced=False, n_samples=10)

src/dataloader.py:324

↓ 1 callersMethodfind_median

(self, train_loader)

src/trainer.py:64

↓ 1 callersFunctionget_flatten_for_nested_list

Return a flatten version of the nested_list contains only select_labels, and a position matrix. Params:

src/utils.py:258

↓ 1 callersFunctionget_saved_path

Create a directory to store the experiment Params: - params: a dict contains all hyper-parameters and experimental setti

src/utils.py:609

↓ 1 callersFunctioninit_experiment

Initialize the experiment, save parameters and create a logger Params: - params: a dict contains all hyper-parameters an

src/utils.py:530

↓ 1 callersFunctionmain_cl

(params)

main_CL.py:17

↓ 1 callersFunctionpdist

Compute the L2 distance of all features Params: - e: a feature matrix has dims (num_samples, hidden_dims) -

src/utils.py:193

↓ 1 callersFunctionplot_centers

Plot the feature centers X Params: - X: a feature matrix has dims (num_classes, hidden_dims) - label_list: a

src/utils.py:306

↓ 1 callersFunctionplot_confusion_matrix

Plot confusion matrix for model predictions Params: - pred_list: a tensor has dims (num_samples,) - y_list:

src/utils.py:397

↓ 1 callersFunctionplot_prob_hist_each_class

Plot probability histogram for each class Params: - y_list: a tensor has dims (num_samples,) - logits_list:

src/utils.py:429

↓ 1 callersMethodreset_parameters

(self)

src/model.py:70

↓ 1 callersFunctionspilt_dataset

(datapath, phase, entity_list, nb_class_fg, nb_class_pg, schema)

src/dataloader.py:106

Method__getitem__

(self, index)

src/dataloader.py:302

src/dataloader.py:298

Method__init__

(self, data_path, domain_name, batch_size, entity_list=[], n_samples=-1, is_filter_O=False, schema='BIO', is_l

src/dataloader.py:420

Method__init__

(self, params, model, label_list)

src/trainer.py:31

Method__init__

(self, hidden_dim, output_dim, sigma=True)

src/model.py:59

Method__init__

(self, hidden_dim, old_output_dim, new_output_dim, sigma=True)

src/model.py:102

Method__len__

(self)

src/dataloader.py:305

Functionassert_no_old_samples

Check the labels contains no samples from old classes Params: - labels: a 2-dimentional tensor - ref_dims: t

src/utils.py:514

Functioncollate_fn

(data)

src/dataloader.py:308

Functioncompute_feature_by_input

Compute the feature of X Params: - X: input sentences have dims (num_sentences, num_words) - feature_model:

src/utils.py:168

Functionconvert_BIOES_to_BIO

(in_datapath, out_datapath)

src/dataloader.py:179

Functiondecode_sentence

Decode the sentences batch from ids to words (string) Params: - sentence: a list of ids (encoded by the tokenizer)

src/utils.py:468

Functiondecode_word_from_sentence

Decode the i-th word from sentence Params: - sentence: a list of ids (encoded by the tokenizer) - pos_idx: t

(self, X, return_feat=False)

src/model.py:34

Methodforward

(self, input, num_head=1)

src/model.py:76

Methodforward

(self, x, num_head=1)

src/model.py:115

Functionget_center

Compute the class center of X, Note that this function is suitable for all classes computation, A better implementation will

src/utils.py:36

Functionget_entity_list

(datapth)

src/dataloader.py:196

Functionget_match_id

Compute the nearest samples id for each sample, Params: - flatten_feat_train: a matrix has dims (num_samples, hidden_dim

src/utils.py:215

Functionplot_distribution

Visualize the feature X in the 2-D space Params: - X: a feature matrix has dims (num_samples, hidden_dims) -

src/utils.py:323

Functionprint_split_data_statistic

(datapath, phase, entity_list, nb_class_fg, nb_class_pg, schema)

Functionspilt_dataset_random

(datapath, phase, entity_list, nb_class_fg, nb_class_pg, schema)

src/dataloader.py:56