method_obj = MethodGraphBertNodeClassification(bert_config,
pretrained_path,
dataset_name)
#---- set to false to run faster ----
method_obj.spy_tag = True
method_obj.max_epoch = max_epoch
method_obj.lr = lr
if use_pretrain:
method_obj.save_pretrained_path = './result/PreTrained_GraphBert/' + dataset_name + '/node_pretrain_classification_model/'
else:
method_obj.save_pretrained_path = './result/PreTrained_GraphBert/' + dataset_name + '/node_classification_model/'
result_obj = ResultSaving()
result_obj.result_destination_folder_path = './result/GraphBert/'
result_obj.result_destination_file_name = dataset_name + '_' + str(
num_hidden_layers)
setting_obj = NewSettings()
evaluate_obj = None
# ------------------------------------------------------
# ---- running section ---------------------------------
setting_obj.prepare(data_obj, method_obj, result_obj, evaluate_obj)
setting_obj.load_run_save_evaluate()
# ------------------------------------------------------
method_obj.save_pretrained('./result/PreTrained_GraphBert/' +
dataset_name +
'/node_classification_complete_model/')
print('************ Finish ************')
data_obj = DatasetLoader()
data_obj.dataset_source_folder_path = './data/' + dataset_name + '/'
data_obj.dataset_name = dataset_name
data_obj.k = k
data_obj.load_all_tag = True
bert_config = GraphBertConfig(residual_type = residual_type, k=k, x_size=nfeature, y_size=y_size, hidden_size=hidden_size, intermediate_size=intermediate_size, num_attention_heads=num_attention_heads, num_hidden_layers=num_hidden_layers)
method_obj = MethodGraphBertNodeConstruct(bert_config)
method_obj.max_epoch = max_epoch
method_obj.lr = lr
method_obj.save_pretrained_path = './result/PreTrained_GraphBert/' + dataset_name + '/node_reconstruct_model/'
result_obj = ResultSaving()
result_obj.result_destination_folder_path = './result/GraphBert/'
result_obj.result_destination_file_name = dataset_name + '_' + str(k) + '_node_reconstruction'
setting_obj = Settings()
evaluate_obj = None
# ------------------------------------------------------
# ---- running section ---------------------------------
setting_obj.prepare(data_obj, method_obj, result_obj, evaluate_obj)
setting_obj.load_run_save_evaluate()
# ------------------------------------------------------
# method_obj.save_pretrained('./result/PreTrained_GraphBert/' + dataset_name + '/node_reconstruct_model/')
print('************ Finish ************')
#------------------------------------
#---- Step 1: Load Raw Graphs for Train/Test Partition ----
if 1:
print('************ Start ************')
print('Preprocessing dataset: ' + dataset_name)
# ---- objection initialization setction ---------------
data_obj = DatasetLoader()
data_obj.dataset_source_folder_path = './data/' + dataset_name + '/'
data_obj.dataset_name = dataset_name
data_obj.load_type = 'Raw'
method_obj = MethodProcessRaw()
result_obj = ResultSaving()
result_obj.result_destination_folder_path = './result/Preprocess/'
result_obj.result_destination_file_name = dataset_name
setting_obj = Settings()
evaluate_obj = None
# ------------------------------------------------------
# ---- running section ---------------------------------
setting_obj.prepare(data_obj, method_obj, result_obj, evaluate_obj)
setting_obj.load_run_save_evaluate()
# ------------------------------------------------------
print('************ Finish ************')
#------------------------------------
#---- Step 2: WL based graph coloring ----
#---- isolated_segment, padding_pruning, full_input
strategy = 'isolated_segment'
dataset_name = 'IMDBMULTI'
residual_type = 'none'
if 1:
epoch_number = 500
result_obj = ResultSaving('', '')
result_obj.result_destination_folder_path = './result/AuGBert/' + strategy + '/' + dataset_name + '/'
result_list = []
time_list = []
for fold in range(1, 11):
result_obj.result_destination_file_name = dataset_name + '_' + str(
fold) + '_' + str(
epoch_number) + '_' + residual_type + '_' + strategy
loaded_result = result_obj.load()
time_list.append(
sum([loaded_result[epoch]['time'] for epoch in loaded_result]))
result_list.append(
np.max(
[loaded_result[epoch]['acc_test'] for epoch in loaded_result]))
print('accuracy: {:.2f}$\pm${:.2f}'.format(100 * np.mean(result_list),
100 * np.std(result_list)))
print('time: {:.2f}$\pm${:.2f}'.format(np.mean(time_list),
np.std(time_list)))
dataset_name = 'PROTEINS'
strategy = 'padding_pruning'
intermediate_size=intermediate_size,
num_attention_heads=num_attention_heads,
num_hidden_layers=num_hidden_layers)
method_obj = MethodGraphBertNodeClassification(bert_config)
#---- set to false to run faster ----
method_obj.spy_tag = True
method_obj.max_epoch = max_epoch
method_obj.lr = lr
method_obj.fold = fold
method_obj.strategy = strategy
method_obj.load_pretrained_path = ''
method_obj.save_pretrained_path = ''
result_obj = ResultSaving()
result_obj.result_destination_folder_path = './result/AuGBert/'
result_obj.result_destination_file_name = dataset_name + '_' + str(
fold) + '_' + str(max_epoch) + '_' + residual_type + '_' + strategy
setting_obj = Settings()
evaluate_obj = None
# ------------------------------------------------------
# ---- running section ---------------------------------
setting_obj.prepare(data_obj, method_obj, result_obj, evaluate_obj)
setting_obj.load_run_save_evaluate()
# ------------------------------------------------------
#method_obj.save_pretrained('./result/PreTrained_GraphBert/' + dataset_name + '/node_classification_complete_model/')
print('************ Finish ************')
#------------------------------------
ngraph = 19717
#---- Step 1: WL based graph coloring ----
if 1:
print('************ Start ************')
print('WL, dataset: ' + dataset_name)
# ---- objection initialization setction ---------------
data_obj = DatasetLoader()
data_obj.dataset_source_folder_path = './data/' + dataset_name + '/'
data_obj.dataset_name = dataset_name
method_obj = MethodWLNodeColoring()
result_obj = ResultSaving()
result_obj.result_destination_folder_path = './result/WL/'
result_obj.result_destination_file_name = dataset_name
setting_obj = Settings()
evaluate_obj = None
# ------------------------------------------------------
# ---- running section ---------------------------------
setting_obj.prepare(data_obj, method_obj, result_obj, evaluate_obj)
setting_obj.load_run_save_evaluate()
# ------------------------------------------------------
print('************ Finish ************')
#------------------------------------
#---- Step 2: intimacy calculation and subgraph batching ----
import matplotlib.pyplot as plt
from code.ResultSaving import ResultSaving
#---------- clustering results evaluation -----------------
dataset_name = 'pubmed'
if 0:
pre_train_task = 'node_reconstruction+structure_recovery'
result_obj = ResultSaving('', '')
result_obj.result_destination_folder_path = './result/GraphBert/'
result_obj.result_destination_file_name = 'clustering_' + dataset_name + '_' + pre_train_task
loaded_result = result_obj.load()
eval_obj = EvaluateClustering()
eval_obj.data = loaded_result
eval_result = eval_obj.evaluate()
print(eval_result)
#--------------- Graph Bert Pre-Training Records Convergence --------------
dataset_name = 'cora'
if 0:
if dataset_name == 'cora':
k = 7
elif dataset_name == 'citeseer':
k = 5
elif dataset_name == 'pubmed':