-
Notifications
You must be signed in to change notification settings - Fork 3
/
Copy pathreproduce.py
93 lines (78 loc) · 7.93 KB
/
reproduce.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
from train4tune import main
import argparse
parser = argparse.ArgumentParser("pas")
parser.add_argument('--gpu', type=int, default=0, help='gpu device id')
parser.add_argument('--data', type=str, default='DD', choices=['DD', 'PROTEINS', 'NCI109', 'IMDB-BINARY', 'COX2', 'IMDB-MULTI'])
args1 = parser.parse_args()
DD_params = {'activation': 'relu', 'hidden_size': 32, 'in_dropout': 0.2, 'learning_rate': 0.006878897133613912,
'model': 'SANE', 'optimizer': 'adam', 'out_dropout': 0.1, 'weight_decay': 1.8447814774886348e-05,
'rnd_num': 1, 'ft_mode': '10fold', 'data': 'DD',
'graph_classification_dataset': ['DD', 'MUTAG', 'PROTEINS', 'NCI1', 'NCI109', 'IMDB-BINARY', 'REDDIT-BINARY', 'BZR', 'COX2', 'IMDB-MULTI', 'COLORS-3', 'COLLAB', 'REDDIT-MULTI-5K'],
'node_classification_dataset': ['Cora', 'CiteSeer', 'PubMed', 'Amazon_Computers', 'Coauthor_CS', 'Coauthor_Physics', 'Amazon_Photo'],
'epochs': 100, 'is_mlp': False, 'batch_size': 64, 'arch': 'gcn||gcn||leaky_relu||leaky_relu||mlppool||gappool||global_sum||none||set2set||l_lstm',
'gpu': 5, 'num_layers': 2, 'seed': 2, 'grad_clip': 5, 'momentum': 0.9, 'cos_lr': True, 'lr_min': 0.0, 'BN': False, 'with_linear': True,
'with_layernorm': False, 'with_layernorm_learnable': True, 'show_info': False, 'withoutjk': False, 'search_act': False,
'one_pooling': False, 'remove_pooling': False, 'remove_jk': False, 'remove_readout': False, 'fixpooling': 'null', 'fixjk': False}
PROTEINS_params = {'activation': 'elu', 'hidden_size': 16, 'in_dropout': 0.0, 'learning_rate': 0.007662125400401121,
'model': 'SANE', 'optimizer': 'adam', 'out_dropout': 0.6, 'weight_decay': 5.979931414632729e-05,
'rnd_num': 1, 'ft_mode': '10fold', 'data': 'PROTEINS',
'graph_classification_dataset': ['DD', 'MUTAG', 'PROTEINS', 'NCI1', 'NCI109', 'IMDB-BINARY', 'REDDIT-BINARY', 'BZR', 'COX2', 'IMDB-MULTI'],
'node_classification_dataset': ['Cora', 'CiteSeer', 'PubMed', 'Amazon_Computers', 'Coauthor_CS', 'Coauthor_Physics', 'Amazon_Photo'],
'epochs': 100, 'is_mlp': False, 'batch_size': 64, 'arch': 'mlp||gin||leaky_relu||softplus||hoppool_3||sagpool||global_sum||set2set||set2set||l_max',
'gpu': 6, 'num_layers': 2, 'seed': 2, 'grad_clip': 5, 'momentum': 0.9, 'cos_lr': True, 'lr_min': 0.0, 'BN': False, 'with_linear': True,
'with_layernorm': False, 'with_layernorm_learnable': True, 'show_info': False, 'withoutjk': False, 'search_act': False,
'one_pooling': False, 'remove_pooling': False, 'remove_jk': False, 'remove_readout': False, 'fixpooling': 'null', 'fixjk': False}
IMDBB_params = {'activation': 'relu', 'hidden_size': 128, 'in_dropout': 0, 'learning_rate': 0.0037532981689691056,
'model': 'SANE', 'optimizer': 'adagrad', 'out_dropout': 0, 'weight_decay': 4.647697507807884e-05,
'rnd_num': 1, 'ft_mode': '10fold', 'data': 'IMDB-BINARY',
'graph_classification_dataset': ['DD', 'MUTAG', 'PROTEINS', 'NCI1', 'NCI109', 'IMDB-BINARY', 'REDDIT-BINARY', 'BZR', 'COX2', 'IMDB-MULTI', 'COLORS-3'],
'node_classification_dataset': ['Cora', 'CiteSeer', 'PubMed', 'Amazon_Computers', 'Coauthor_CS', 'Coauthor_Physics', 'Amazon_Photo'],
'epochs': 100, 'is_mlp': False, 'batch_size': 64, 'arch': 'gin||gcn||sigmoid||elu||hoppool_3||mlppool||global_sort||global_mean||global_att||l_sum',
'gpu': 6, 'num_layers': 2, 'seed': 2, 'grad_clip': 5, 'momentum': 0.9, 'cos_lr': True, 'lr_min': 0.0, 'BN': False, 'with_linear': True,
'with_layernorm': False, 'with_layernorm_learnable': True, 'show_info': False, 'withoutjk': False, 'search_act': False,
'one_pooling': False, 'remove_pooling': False, 'remove_jk': False, 'remove_readout': False, 'fixpooling': 'null', 'fixjk': False}
IMDBM_params = {'activation': 'elu', 'hidden_size': 64, 'in_dropout': 0.0, 'learning_rate': 0.008736477470973399,
'model': 'SANE', 'optimizer': 'adagrad', 'out_dropout': 0.0, 'weight_decay': 0.0004603406614944098,
'rnd_num': 1, 'ft_mode': '10fold', 'data': 'IMDB-MULTI',
'graph_classification_dataset': ['DD', 'MUTAG', 'PROTEINS', 'NCI1', 'NCI109', 'IMDB-BINARY', 'REDDIT-BINARY', 'BZR', 'COX2', 'IMDB-MULTI'],
'node_classification_dataset': ['Cora', 'CiteSeer', 'PubMed', 'Amazon_Computers', 'Coauthor_CS', 'Coauthor_Physics', 'Amazon_Photo'],
'epochs': 100, 'is_mlp': False, 'batch_size': 64, 'arch': 'gat||relu6||sag_graphconv||global_sum||global_mean||l_lstm',
'gpu': 7, 'num_layers': 1, 'seed': 2, 'grad_clip': 5, 'momentum': 0.9, 'cos_lr': True, 'lr_min': 0.0, 'BN': False, 'with_linear': True,
'with_layernorm': False, 'with_layernorm_learnable': True, 'show_info': False, 'withoutjk': False, 'search_act': False,
'one_pooling': False, 'remove_pooling': False, 'remove_jk': False, 'remove_readout': False, 'fixpooling': 'null', 'fixjk': False}
COX2_patams = {'activation': 'relu', 'hidden_size': 16, 'in_dropout': 0.0, 'learning_rate': 0.001921018166741771,
'model': 'SANE', 'optimizer': 'adam', 'out_dropout': 0.2, 'weight_decay': 6.593806454834699e-05,
'rnd_num': 1, 'ft_mode': '10fold', 'data': 'COX2',
'graph_classification_dataset': ['DD', 'MUTAG', 'PROTEINS', 'NCI1', 'NCI109', 'IMDB-BINARY', 'REDDIT-BINARY', 'BZR', 'COX2', 'IMDB-MULTI'],
'node_classification_dataset': ['Cora', 'CiteSeer', 'PubMed', 'Amazon_Computers', 'Coauthor_CS', 'Coauthor_Physics', 'Amazon_Photo'],
'epochs': 140, 'is_mlp': False, 'batch_size': 64, 'arch': 'gin||gat||relu||tanh||none||gappool||global_sum||none||global_att||l_lstm',
'gpu': 7, 'num_layers': 2, 'seed': 2, 'grad_clip': 5, 'momentum': 0.9, 'cos_lr': False, 'BN': False, 'lr_min': 0.0, 'with_linear': False,
'with_layernorm': True, 'with_layernorm_learnable': False,'show_info': False, 'withoutjk': False, 'search_act': False,
'one_pooling': False, 'remove_pooling': False, 'remove_jk': False, 'remove_readout': False, 'fixpooling': 'null', 'fixjk': False}
NCI109_params = {'activation': 'relu', 'hidden_size': 32, 'in_dropout': 0.1, 'learning_rate': 0.005385412452868308,
'model': 'SANE', 'optimizer': 'adam', 'out_dropout': 0.1, 'weight_decay': 1.2956857470096877e-05,
'rnd_num': 1, 'ft_mode': '10fold', 'data': 'NCI109',
'graph_classification_dataset': ['DD', 'MUTAG', 'PROTEINS', 'NCI1', 'NCI109', 'IMDB-BINARY', 'REDDIT-BINARY', 'BZR', 'COX2', 'IMDB-MULTI', 'COLORS-3', 'COLLAB', 'REDDIT-MULTI-5K'],
'node_classification_dataset': ['Cora', 'CiteSeer', 'PubMed', 'Amazon_Computers', 'Coauthor_CS', 'Coauthor_Physics', 'Amazon_Photo'],
'epochs': 100, 'is_mlp': False, 'batch_size': 128, 'arch': 'graphconv_add||gin||relu6||leaky_relu||hoppool_1||mlppool||global_sum||set2set||global_att||l_lstm',
'gpu': 3, 'num_layers': 2, 'seed': 2, 'grad_clip': 5, 'momentum': 0.9, 'cos_lr': True, 'lr_min': 0.0, 'BN': False, 'with_linear': True,
'with_layernorm': False, 'with_layernorm_learnable': True, 'show_info': False, 'withoutjk': False, 'search_act': False,
'one_pooling': False, 'remove_pooling': False, 'remove_jk': False, 'remove_readout': False, 'fixpooling': 'null', 'fixjk': False}
params_dict = {
'DD': DD_params,
'PROTEINS': PROTEINS_params,
'NCI109': NCI109_params,
'COX2': COX2_patams,
'IMDB-BINARY': IMDBB_params,
'IMDB-MULTI': IMDBM_params
}
class Dict(dict):
__setattr__ = dict.__setitem__
__getattr__ = dict.__getitem__
args = Dict()
params_dict[args1.data]['gpu'] = args1.gpu
for k, v in params_dict[args1.data].items():
args[k] = v
for i in range(5):
valid_acc, test_acc, test_std, args = main(args)
print('{}/5: valid_acc:{:.04f}, test_acc: {:.04f}+-{:.04f}'.format(i, valid_acc, test_acc, test_std))