Functions — github.com/chengsen/PyTorch_TextGCN

CodeHub MCPcopy Create free account

hub / github.com/chengsen/PyTorch_TextGCN / functions

Functions61 in github.com/chengsen/PyTorch_TextGCN

⨍ Functions 61 ◇ Types & classes 10

↓ 3 callersFunctionprint_graph_detail

格式化显示Graph参数 :param graph: :return:

↓ 2 callersMethodclean_str

data_processor.py:20

↓ 2 callersMethodforward

↓ 2 callersMethodlog

(self, result: dict)

↓ 2 callersMethodremove_stopword

data_processor.py:67

↓ 2 callersMethodsave

build_graph.py:177

↓ 2 callersMethodval

(self, x, prefix="val")

↓ 1 callersMethod__init__

(self, nfeat, nhid, nclass, dropout)

↓ 1 callersFunctionaccuracy

↓ 1 callersFunctioncal_pmi

(W_ij, W, word_freq_1, word_freq_2)

build_graph.py:54

↓ 1 callersMethodclean_text

data_processor.py:112

↓ 1 callersMethodconvert_tensor

↓ 1 callersFunctioncount_pmi

(windows_len, word_pair_count, word_window_freq, threshold)

build_graph.py:63

↓ 1 callersMethodfit

↓ 1 callersFunctionget_pmi_edge

(content_lst, window_size=20, threshold=0.)

build_graph.py:76

↓ 1 callersMethodget_pmi_edge

build_graph.py:110

↓ 1 callersMethodget_tfidf_edge

build_graph.py:123

↓ 1 callersMethodget_tfidf_vec

学习获得tfidf矩阵，及其对应的单词序列 :param content_lst: :return:

build_graph.py:139

↓ 1 callersFunctionget_train_test

↓ 1 callersFunctionget_window

找出窗口 :param content_lst: :param window_size: :return:

build_graph.py:19

↓ 1 callersMethodgpu_mem_get

get total, used and free memory (in MBs) for gpu `id`. if `id` is not passed, currently selected torch device is used

↓ 1 callersMethodgpu_mem_get_all

get total, used and free memory (in MBs) for each available gpu

↓ 1 callersMethodlog_single

(self, key, value)

↓ 1 callersFunctionmacro_f1

(pred, targ, num_classes=None)

↓ 1 callersFunctionmain

build_graph.py:185

↓ 1 callersFunctionmain

data_processor.py:152

↓ 1 callersFunctionmain

(dataset, times)

↓ 1 callersFunctionnormalize_adj

Symmetrically normalize adjacency matrix.

↓ 1 callersFunctionparameter_parser

A method to parse up command line parameters. By default it trains on the PubMed yelp_dataset. The default hyperparameters give a good qualit

↓ 1 callersMethodprepare_data

↓ 1 callersFunctionpreprocess_adj

Preprocessing of adjacency matrix for simple pygGCN model and conversion to tuple representation.

↓ 1 callersFunctionprint_table

↓ 1 callersFunctionread_file

(path, mode='r', encoding=None)

↓ 1 callersFunctionremove_less_word

(lines_str, word_st)

data_processor.py:93

↓ 1 callersMethodreset_parameters

↓ 1 callersFunctionreturn_seed

↓ 1 callersMethodset_description

↓ 1 callersMethodset_seed

↓ 1 callersMethodshow_str

↓ 1 callersFunctionsparse_mx_to_torch_sparse_tensor

Convert a scipy sparse matrix to a torch sparse tensor.

↓ 1 callersMethodtest

↓ 1 callersMethodtrain

(self, val_loss, model=None)

(self, dataset)

build_graph.py:92

data_processor.py:11

(self, dataset, encoding=None)

data_processor.py:98

Args: patience (int): How long to wait after last time validation loss improved. Default: 7

(self, in_features, out_features, bias=True)

(self, args, model, pre_data)

(self, infeatn, adj)

Methodget_cuda_id

Methodlean_str_sst

Tokenization/string cleaning for the SST yelp_dataset Original taken from https://github.com/yoonkim/CNN_sentence/blob/master

data_processor.py:58

Methodload_model

data_processor.py:37

Methodreplace_num

data_processor.py:83

Methodreplace_urls

data_processor.py:87

Methodsave_checkpoint

Saves model when validation loss decrease.