-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathparse.py
229 lines (212 loc) · 12.7 KB
/
parse.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
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
from models import *
from ours import *
from nodeformer import *
from graphgps import *
def parse_method(method, args, c, d, device):
if method == 'gcn':
model = GCN(in_channels=d,
hidden_channels=args.hidden_channels,
out_channels=c,
num_layers=args.num_layers,
dropout=args.dropout,
use_bn=args.use_bn).to(device)
elif method == 'sgc':
model = SGC(in_channels=d,
out_channels=c,
hops=args.hops).to(device)
elif method == 'gat':
model = GAT(in_channels=d,
hidden_channels=args.hidden_channels,
out_channels=c,
num_layers=args.num_layers,
dropout=args.dropout,
use_bn=args.use_bn,
heads=args.gat_heads,
out_heads=args.out_heads).to(device)
elif method == 'gcnjk':
model = GCNJK(in_channels=d,
hidden_channels=args.hidden_channels,
out_channels=c,
num_layers=args.num_layers,
dropout=args.dropout).to(device)
elif method == 'appnp':
model = APPNP_Net(in_channels=d,
hidden_channels=args.hidden_channels,
out_channels=c,
dropout=args.dropout).to(device)
elif method == 'h2gcn':
model = model = H2GCN(feat_dim=d,hidden_dim=args.hidden_channels,class_dim=c,dropout=args.dropout).to(device)
elif method == 'sign':
model = SIGN(in_channels=d,
hidden_channels=args.hidden_channels,
out_channels=c,
hops=args.hops,
num_layers=args.num_layers,
dropout=args.dropout,
use_bn=args.use_bn).to(device)
elif method == 'nodeformer':
model = NodeFormer(in_channels=d,
hidden_channels=args.hidden_channels,
out_channels=c,
num_layers=args.num_layers,
dropout=args.dropout,
num_heads=args.num_heads,
use_bn=args.use_bn).to(device)
elif method == 'ours':
if args.use_graph:
gnn=parse_method(args.backbone, args, args.hidden_channels, d, device)
model = SGFormer(d, args.hidden_channels, c, num_layers=args.ours_layers, alpha=args.alpha, dropout=args.ours_dropout, num_heads=args.num_heads,
use_bn=args.use_bn, use_residual=args.ours_use_residual, use_graph=args.use_graph, use_weight=args.ours_use_weight, use_act=args.ours_use_act, graph_weight=args.graph_weight, gnn=gnn, aggregate=args.aggregate).to(device)
elif args.method == 'sgformer':
model = SGFormer_large(d, args.hidden_channels, c, graph_weight=args.graph_weight, aggregate=args.aggregate,
trans_num_layers=args.trans_num_layers, trans_dropout=args.trans_dropout, trans_num_heads=args.trans_num_heads, trans_use_bn=args.trans_use_bn, trans_use_residual=args.trans_use_residual, trans_use_weight=args.trans_use_weight, trans_use_act=args.trans_use_act,
gnn_num_layers=args.gnn_num_layers, gnn_dropout=args.gnn_dropout, gnn_use_bn=args.gnn_use_bn, gnn_use_residual=args.gnn_use_residual, gnn_use_weight=args.gnn_use_weight, gnn_use_init=args.gnn_use_init, gnn_use_act=args.gnn_use_act,
).to(device)
elif method == 'graphgps':
model = GPSModel(in_channels=d,
out_channels=c,
hidden_channels=args.hidden_channels,
num_layers=args.num_layers,
num_heads=args.num_heads,
dropout=args.dropout,
attn_dropout=args.dropout,
use_bn=args.use_bn).to(device)
elif method == 'nodeformer_sub':
model = NodeFormerLarge(in_channels=d,
hidden_channels=args.hidden_channels,
out_channels=c,
num_layers=args.num_layers,
dropout=args.dropout,
num_heads=args.num_heads,
use_bn=args.use_bn).to(device)
else:
raise ValueError(f'Invalid method {method}')
return model
def parser_add_main_args(parser):
# dataset and evaluation
# parser.add_argument('--data_dir', type=str, default='../../../NodeFormer/data/')
parser.add_argument('--data_dir', type=str, default='../../data')
parser.add_argument('--dataset', type=str, default='cora')
parser.add_argument('--device', type=int, default=0,
help='which gpu to use if any (default: 0)')
parser.add_argument('--seed', type=int, default=42)
parser.add_argument('--cpu', action='store_true')
parser.add_argument('--epochs', type=int, default=500)
parser.add_argument('--runs', type=int, default=1,
help='number of distinct runs')
parser.add_argument('--train_prop', type=float, default=.5,
help='training label proportion')
parser.add_argument('--valid_prop', type=float, default=.25,
help='validation label proportion')
parser.add_argument('--protocol', type=str, default='semi',
help='protocol for cora datasets, semi or supervised')
parser.add_argument('--rand_split', action='store_true',
help='use random splits')
parser.add_argument('--rand_split_class', action='store_true',
help='use random splits with a fixed number of labeled nodes for each class')
parser.add_argument('--label_num_per_class', type=int, default=20,
help='labeled nodes per class(randomly selected)')
parser.add_argument('--valid_num', type=int, default=500,
help='Total number of validation')
parser.add_argument('--test_num', type=int, default=500,
help='Total number of test')
parser.add_argument('--directed', action='store_true',
help='set to not symmetrize adjacency')
# Add metric
parser.add_argument('--metric', type=str, default='acc', choices=['acc', 'rocauc', 'f1'],
help='evaluation metric')
# model
parser.add_argument('--method', type=str, default='gcn')
parser.add_argument('--hidden_channels', type=int, default=32)
parser.add_argument('--num_layers', type=int, default=2,
help='number of layers for deep methods')
parser.add_argument('--num_heads', type=int, default=1,
help='number of heads for attention')
parser.add_argument('--alpha', type=float, default=0.5,
help='weight for residual link')
parser.add_argument('--use_bn', action='store_true', help='use layernorm')
parser.add_argument('--use_residual', action='store_true',
help='use residual link for each GNN layer')
parser.add_argument('--use_graph', action='store_true', help='use pos emb')
parser.add_argument('--use_weight', action='store_true',
help='use weight for GNN convolution')
parser.add_argument('--use_init', action='store_true', help='use initial feat for each GNN layer')
parser.add_argument('--use_act', action='store_true', help='use activation for each GNN layer')
parser.add_argument('--attention', type=str, default='gcn')
# training
parser.add_argument('--lr', type=float, default=0.01)
parser.add_argument('--weight_decay', type=float, default=5e-3)
parser.add_argument('--batch_size', type=int, default=10000, help='mini batch training for large graphs')
parser.add_argument('--dropout', type=float, default=0.5)
# display and utility
parser.add_argument('--display_step', type=int,
default=50, help='how often to print')
parser.add_argument('--eval_step', type=int,
default=1, help='how often to evaluate')
parser.add_argument('--cached', action='store_true',
help='set to use faster sgc')
parser.add_argument('--print_prop', action='store_true',
help='print proportions of predicted class')
parser.add_argument('--save_result', action='store_true',
help='save result')
parser.add_argument('--save_model', action='store_true', help='whether to save model')
parser.add_argument('--use_pretrained', action='store_true', help='whether to use pretrained model')
parser.add_argument('--save_att', action='store_true', help='whether to save attention (for visualization)')
parser.add_argument('--model_dir', type=str, default='../../model/')
parser.add_argument('--no_feat_norm', action='store_true',
help='Not use feature normalization.')
# ours
parser.add_argument('--patience', type=int, default=200,
help='early stopping patience.')
parser.add_argument('--graph_weight', type=float,
default=0.8, help='graph weight.')
parser.add_argument('--ours_weight_decay', type=float,
help='Ours\' weight decay. Default to weight_decay.')
parser.add_argument('--ours_use_weight', action='store_true', help='use weight for trans convolution')
parser.add_argument('--ours_use_residual', action='store_true', help='use residual link for each trans layer')
parser.add_argument('--ours_use_act', action='store_true', help='use activation for each trans layer')
parser.add_argument('--backbone', type=str, default='gcn',
help='Backbone.')
parser.add_argument('--ours_layers', type=int, default=2,
help='gnn layer.')
parser.add_argument('--ours_dropout', type=float,
help='gnn dropout.')
parser.add_argument('--aggregate', type=str, default='add',
help='aggregate type, add or cat.')
# hyper-parameter for gnn baseline
parser.add_argument('--hops', type=int, default=1,
help='power of adjacency matrix for certain methods')
parser.add_argument('--gat_heads', type=int, default=8,
help='attention heads for gat')
parser.add_argument('--out_heads', type=int, default=1,
help='out heads for gat')
parser.add_argument('--lamda', type=float, default=0.1, help='weight for edge reg loss')
# graphormer
parser.add_argument('--num_elayers', type=int, default=2,
help='number of encoder layers for graphormers')
parser.add_argument('--encoder_emdim', type=int, default=768,
help='number of encoder embedded dimension')
# GNN Branch
parser.add_argument('--gnn_use_bn', action='store_true', help='use batchnorm for each GNN layer')
parser.add_argument('--gnn_use_residual', action='store_true', help='use residual link for each GNN layer')
parser.add_argument('--gnn_use_weight', action='store_true', help='use weight for GNN convolution')
parser.add_argument('--gnn_use_init', action='store_true', help='use initial feat for each GNN layer')
parser.add_argument('--gnn_use_act', action='store_true', help='use activation for each GNN layer')
parser.add_argument('--gnn_num_layers', type=int, default=2, help='number of layers for GNN')
parser.add_argument('--gnn_dropout', type=float, default=0.0)
parser.add_argument('--gnn_weight_decay', type=float, default=1e-3)
# all-pair attention (Transformer) branch
parser.add_argument('--trans_num_heads', type=int, default=1, help='number of heads for attention')
parser.add_argument('--trans_use_weight', action='store_true', help='use weight for trans convolution')
parser.add_argument('--trans_use_bn', action='store_true', help='use layernorm for trans')
parser.add_argument('--trans_use_residual', action='store_true', help='use residual link for each trans layer')
parser.add_argument('--trans_use_act', action='store_true', help='use activation for each trans layer')
parser.add_argument('--trans_num_layers', type=int, default=2, help='number of layers for all-pair attention.')
parser.add_argument('--trans_dropout', type=float, help='gnn dropout.')
parser.add_argument('--trans_weight_decay', type=float, default=1e-3)
def parser_add_default_args(args):
if args.method=='ours':
if args.ours_weight_decay is None:
args.ours_weight_decay=args.weight_decay
if args.ours_dropout is None:
args.ours_dropout=args.dropout